Interesting and useful articles about marketing, our product and online communications

Tips for Digital Marketing & Sales in Altcraft Blog

We’ll send you a guide 

with 26 strategies for
automating marketing via email.

Our experts will guide you in maximizing the effectiveness of Altcraft Platform for achieving your goals.

Elevate Your Business 
to a New Level

By continuing to use our site, you consent to the processing of cookies that ensure the proper functioning of the site.

Download documentation

Thank you!

Feel free to ask us any questions. Our experts are


Ask our expert 

ready to answer them during business hours.

Anti-Spam policy

Blog

Contacts

Support

Partners

Pricing

Privacy Policy

Why Altcraft

Page not found

Help Center: Guides on How to Use Altcraft Platform

Personal data policy

Privacy Policy
and personal data policy

Request a demo | Fill Out the Form | Altcraft Platform

Get Started With Altcraft's Software

Why business chooses Altcraft

processing of personal data

You can ask us your questions. Our experts are ready to answer them within the working day.

What else 
would you like to know?

How to set up a tracking pixel

What is a loyalty program

How to create a customer database

How to manually import customer data

You may be interested

<Video
url="https://www.youtube.com/embed/KoB1iuxODE8?si=gvegnrGqBNJ9RmNa"
/>

The effectiveness of mobile applications and online services is closely tied to user activity. If people are logging in and using the service, it indicates that the developers' idea is successful and will generate revenue. It is important to track user behavior and conduct analytics to understand the problems and growth areas of the product.

This issue is addressed by product metrics measuring activity over a day (DAU), a week (WAU), and a month (MAU). The time tracking interval is chosen depending on the type of application and the tasks it solves. DAU, WAU, and MAU metrics reflect how popular the product and individual solutions are.

In this article, we will discuss why it is important to track how many users remain active over different time intervals. We will explain the correct calculation of results and ways to influence them.

## DAU metric

The DAU indicator denotes how many active users logged in and/or performed an action in the application in one day (Daily Active Users). Only unique visitors per day are counted. If one person opened the product multiple times, only the first entry is considered.

*Example:*
*300 people downloaded an app to keep a diary. On the second day, 150 of them logged in. Therefore, DAU = 150. Even though some used the app multiple times, only the first interaction is counted.*

### When to apply the DAU metric

Counting active users per day is suitable when the application or service is designed for daily interaction: calendars, planners, habit trackers, games, music, and so on. 
## WAU metric


The WAU indicator reflects the product's weekly active users (Weekly Active Users). Similar to the daily activity metric, only unique visits are counted, but WAU is not the sum of DAU over seven days.

*Example:*
*A banking app was installed by 200 people, and 70 of them logged in at least once during the week. Therefore, WAU is 70.*

### When to apply the WAU metric

This metric is relevant for services/apps intended for regular use, but not necessarily daily. These include health trackers, delivery services, messengers, financial tools, and others.

## MAU metric

The MAU indicator counts users who accessed the product at least once a month (Monthly Active Users). This metric is calculated separately and is not the sum of DAU or WAU.

*Example:*
*A flight booking app was downloaded by 1000 people. Within thirty days, 300 users logged in once, and another 150 logged in three or more times. However, for MAU, only the first login is counted, so the metric is 450.*

### When to apply the MAU metric

MAU is relevant for platforms that do not require daily or even weekly usage. These include services for booking travel, paying taxes, utilities, marketplaces, and so on.

## How to accurately calculate MAU, WAU, DAU

While product metrics do not require complex formulas, there are nuances in their calculation that should be considered.

**Define active users in advance.** It’s essential to specify who counts as an active user. Simply opening an app may not indicate effectiveness, especially in services where specific actions matter. For instance, a user might click on a marketplace app and immediately exit. In such cases, the activity should be defined as viewing products, adding items to the cart, or making a purchase.

**Determine the appropriate time Interval.** Choose the right time interval to track activity. Daily DAU metrics are not relevant for services used infrequently and might incorrectly suggest inefficiency. For example, it would be unreasonable to expect users to log into a fitness club app every day if it mainly functions for class bookings, as few people attend daily workouts.

**Consider external factors in calculations.** Account for external influences when calculating metrics. Marketing campaigns aimed at attracting new users can temporarily boost MAU, WAU, and DAU, which might drop after the campaign ends. This doesn’t necessarily indicate a decline in service quality. News events can also impact user activity, such as reports of widespread food poisoning potentially reducing orders through food delivery apps.

[**Segment your audience.**](https://altcraft.com/glossary/market-segmentation-what-it-is-its-types-and-main-stages) Understanding what attracts new users versus long-term users is crucial. Their needs may differ significantly. Additionally, segment users by time zones to accurately track daily activity (especially for DAU), by devices, and by payment types (one-time purchases, subscriptions, etc.).


## Why calculate MAU, WAU, DAU

These metrics are crucial for understanding the real popularity of a service or application. The number of downloads on the marketplace alone doesn't provide a complete picture, as users might not even open the product after downloading it.

Therefore, DAU, WAU, and MAU are calculated to:

**Monitor the proper functioning of the application/service.** Reduced activity can sometimes be linked to system glitches and errors. Tracking these metrics allows for quick detection and resolution of such issues.

**Evaluate marketing communications and advertising.** Tracking these metrics helps assess how well marketing efforts attract [**the target audience**](https://altcraft.com/glossary/target-audience-why-to-define-it-how-to-classify-and-segment-it), rather than just random users. It also helps measure the effectiveness of specific campaigns aimed at increasing activity and engagement, which is vital for planning future budgets and advertising strategies.

**Test new features.** User activity metrics reveal whether updates are relevant. They help identify which features should be improved or removed to enhance user experience.

**Attract investments.** Stable and growing user activity demonstrates the product's success. For investors, this is a strong incentive to invest, as it indicates a high likelihood of profitability and return on investment.

**Stay relevant in a competitive environment.** Tracking daily, weekly, and monthly activity metrics, and comparing them with competitors, highlights the need for changes or improvements to well-performing features.

## How to improve DAU, WAU, and MAU

Metrics are not static; they are influenced by various factors. Therefore, creators of digital products can adjust these metrics through marketing tools and technical improvements.

Key factors for improving metrics:

**User-friendly interface and attractive design.** Your product should stand out on the user's screen and be intuitively understandable. Complex action scenarios that require a lot of time will not motivate users to return and stay active.

<Img
img="/uploads/attractive_interface_of_netflix_home_page_df0b22d410.png"
alt="Example of attractive interface of Netflix home page"
src="Attractive interface of Netflix home page"
/>

**Stable product performance.** Constant errors, crashes, and slow customer support responses will reduce user activity as they visit applications/services less frequently. Therefore, it is essential to ensure everything works correctly, respond quickly, and resolve issues promptly.

**Content diversity and personalization.** In today's market, with numerous offerings, it's crucial to study user interests in detail, monitor their changes, and adapt accordingly. Don't be afraid to be creative with your content and analyze audience preferences to make precise proposals. This way, users will engage with your products more often.

<Note>
To learn more about content personalization, read the article [**Personalization: What It Is and Why Personalized Marketing Matters**](https://altcraft.com/glossary/personalization-what-it-is-and-why-personalized-marketing-matters).
</Note>

An effective way to diversify content is by introducing gamification elements and a system to reward user activity. 

**Consistent communication with the audience through push and email notifications.** Keep users informed about promotions, discounts, new features, and interactive events in your application or service. Continuing to engage with users within the app is also crucial. Therefore, onboarding and a well-structured communication system are necessary for ongoing interaction.

<Img
img="/uploads/email_promotion_from_sephora_1faa5a73ff.jpg"
alt="Example of email promotion from Sephora beauty store"
src="Email promotion from Sephora beauty store"
/>


**Conduct analytics and respond to user feedback.** Listen to customer reviews to continuously improve your product. Conduct audience preference research, such as custdev, A/B testing, and other methods.

It's important to assess the effectiveness of your service or application from various angles, not just by focusing on WAU, MAU, and DAU. Typically, additional metrics are considered, such as the Sticky Factor, LTV (Lifetime Value, the total revenue from a customer over their lifetime), conversion rates, CAC (Customer Acquisition Cost), and others.

## Summary

MAU, WAU, and DAU are metrics that measure user activity over different time periods: thirty or thirty-one days, seven days, and one day. They are used to assess the popularity of an app or online service, evaluate marketing effectiveness, gauge the impact of new features, and attract investment.

To calculate these metrics, it’s important to define what constitutes user activity, choose the right tracking period (day, week, month), consider external factors, and segment the audience.

To increase MAU, WAU, and DAU, strategies such as targeted messaging, personalization, engaging content, and continuous interface improvements, along with ongoing analytics, are employed.


MAU, WAU, and DAU track user activity to gauge app success. Boost them with targeted messaging, personalization, and engaging content.

metrics-mau-wau-dau.png

About why to track how many users remain active at different time intervals, as well as about the correct calculation of results and ways to influence them.

MAU, WAU, DAU: When to Apply & How to Improve These Metrics

Metrics MAU, WAU, DAU: When to Apply and How to Improve

<Video
url="https://www.youtube.com/embed/SaEo6Iian4E?si=8vOojkN4egDq8IWG"
/>

The goal of any business is not only to attract but also to retain customers. A loyal audience that consistently returns for repeat purchases, renews subscriptions in apps and services, provides companies with more stability, and saves advertising budgets.

Therefore, tracking user retention regularly is essential. In marketing, there is a metric called Retention Rate (RR). This metric considers the ratio of old and new customers over a specific period and indicates when a customer retention strategy needs adjustment.

In this article, you will learn what Retention Rate means, why it is used, how to calculate the customer retention rate, and how to increase RR.
## What is Retention Rate and why measure it

**Retention rate** is a customer retention metric that shows what percentage of people return and use a business's product again. This metric is especially popular in the IT sector, where users need to regularly log into applications and services to benefit their creators.

### Why calculate customer Retention Rate:

**Understand product effectiveness:** Determine how well your product actually solves customer problems so that they continue using it.

**Evaluate marketing and service efficiency:** Assess the effectiveness of marketing departments and the service level provided by the company. A low retention rate is a warning signal. Analyzing it will show which department is underperforming and why customers are not returning.

**Reduce marketing costs:** Retaining customers is cheaper than acquiring new ones, which usually involves multiple advertising campaigns. The retention rate metric shows how effectively your marketing works on existing customers.

**Increase company revenue:** The usefulness and convenience of your product enhance customer loyalty. Loyal customers not only return and respond to additional offers but also recommend your brand to others.

<Note>
In the article "[**Customer Loyalty: What Is It and How to Calculate It**](https://altcraft.com/blog/customer-loyalty)" we explained in detail why loyalty is important and which metrics are used to measure it.
</Note>

## How to calculate Retention Rate

To calculate the customer retention rate, you need the following data:

- **A.** The number of customers at the beginning of the selected calculation period. <p>
</p>
- **B.** New users during this time. <p>
</p>
- **C.** The number of people who continue to use your product at the end of the period.

The Retention Rate formula looks like this:

<Img
img="/uploads/calculate_retention_rate_9b1ea036d4.jpg"
alt="Calculate Retention Rate formula"
/>

The calculation period is chosen depending on the business sector and the product's characteristics. If the purchase cycle is quick, the Retention Rate is calculated more frequently. The usual calculation periods for this metric range from daily to yearly.

***Example RR Calculation #1:***

*At the beginning of the month, the interior design app had 5,679 users.
During the month, an additional 120 users joined. At the end of the month, 5,456 people were using the product.*

(5456 - 120) / 5679 * 100%

*RR = 93%*

***Example RR Calculation #2:***

*Over the year, 11,320 people made purchases in the online store, starting with 13,780 customers at the beginning of the year, and 2,600 new buyers during this period.*

(11320 - 2600) / 13780 * 100%

*RR = 63%*

It is also permissible to calculate the retention rate for separate cohorts of users to ensure proper segmentation and enhance [**the personalization of offers**](https://altcraft.com/glossary/personalization-what-it-is-and-why-personalized-marketing-matters).

## How to evaluate Retention Rate

There is no specific benchmark percentage for RR because each industry has its own figures, and customer acquisition and retention are done differently. Therefore, it is advisable to benchmark against market leaders you aspire to match when calculating your RR. It is important to consider successful cases.

*For example, a retention rate of 65% for a mobile app would be low, whereas it would be high for the retail sector, where* [***the average is 63%***](https://www.paddle.com/resources/industry-retention-rates).

Another way to find your benchmark for RR is by considering your own data from previous periods. By continuously tracking the rate, you will establish your internal ideal percentage. This method is also suitable when competitor RR information is not publicly available.

## Retention Rate and other metrics

Marketing is a system that should not be evaluated based on a single metric. To assess whether your strategy is working when launching campaigns for customer retention and loyalty enhancement, other metrics are also used.
 
<Table>
|**Metric**|**Description**|
| ------ | ------ | 
| **Churn rate** | The number of users (in %) who unsubscribed from emails or newsletters via messengers, SMS, Push and stopped using your service. That is, churn rate metric indicates how many people have stopped connecting with the brand through all communication channels. Churn is a normal process, because products sometimes lose relevance for customers. But it is important that this indicator is overlapped with new users.  |
| **Lifetime value** | Profit from the customer over the entire time they have interacted with the company. A useful metric for personalization because over a long period of time with a customer, the brand gathers enough information about their needs, tastes and orders to make a customized offer. |
| **Redemption Rate** | Metric for evaluating the use of provided bonuses: discount coupons, points and so on. Needed to check the effectiveness of loyalty programs and marketing campaigns.  |
| **Customer Satisfaction Index** | Measuring customer satisfaction with specific interactions with the company and product. It is checked, for example, to what extent users are satisfied with the quality of goods, variety of assortment, communication by phone with service, technical support, and so on. |
| **Repeat Customer Rate** | Indicates how many people out of all customers have returned for purchases two times or more. A high return rate indicates satisfaction with a product or service and its usefulness to users.  |
| **Net Promoter Score** | A metric of user commitment to your company. Shows how willing customers are to buy again and recommend you to others. Users rate the brand on a given scale, and an overall coefficient is calculated for all the answers of the questionnaire.  |
</Table>


## How to improve Retention Rate

To impact RR, it’s crucial to understand what user retention means for your company and what tools and methods are used. Then, choose the available strategies.

**Improve the product.** Increase its value for customers by adding new features, improving existing ones, fixing bugs, and so on. The faster and more effectively your product solves the customer’s problem, the more often they will want to return to it.

**Simplify product use.** Eliminate lengthy and confusing registrations, provide a simple [**onboarding process**](https://altcraft.com/glossary/why-do-you-need-onboarding), and ensure an intuitive interface to enhance the user experience.

**Launch marketing campaigns as retention methods.** Remind users about your product through email newsletters, push notifications, [**messages in messengers**](https://altcraft.com/blog/messenger-marketing), and SMS. Use special offers and promotions to encourage users to return to purchases and other target actions, with personalization being particularly effective. Providing bonuses for various activities, which also benefit your brand, can be useful for retention.

*For example, Nook Vibrant Kitchen, a plant-based meal delivery service, excels in ecommerce email marketing with their loyalty program explainer email. It outlines earning and redeeming points, referrals, and the bottle return program, ending with a join call. This thorough guide helps eliminate customer confusion.*

<Img
img="/uploads/customer_loyalty_program_74bbe1728f.jpg"
alt="Customer loyalty program based on clients’ activities"
src="Nook Vibrant Kitchen gives points for every type of activity"
/>

**Expand the audience and attract new users for whom your product has great value.** Here, the marketing strategy primarily ensures an influx of new users so that the number of loyal and engaged users outweighs those who have lost interest in your product.

**Engage in dialogue with customers.** Ask users for their opinions about the service or product they purchased, and about the ordering, delivery, and other processes. Loyalty increases when people understand that brands listen to them and respond to criticism. Additionally, responses in surveys and questionnaires will point out the strengths and weaknesses of the product and service. Addressing these deficiencies will lead to an increase in Retention Rate.

<Img
img="/uploads/survey_to_ask_for_feedback_4ae6f3f5ad.jpg"
alt="Survey to ask for feedback"
src="The user is asked to take a survey and leave an opinion about the online store's performance"
/>

## Summary

Retention rate is the customer retention rate, calculated over a chosen period of time (day, week, month, etc.). This metric is essential for evaluating the product's usefulness to the audience, reducing marketing expenses, and increasing business sales.

The RR formula considers the ratio of customers at the beginning and end of the period, along with the number of new users. There is no universal rule for what the customer retention rate should be; for evaluation, examples from one's own business field or previous data are used.

In addition to Retention Rate, other metrics can be calculated to increase marketing effectiveness, such as customer lifetime value, churn rate, repeat purchases, use of rewards, and loyalty and satisfaction indexes.

To improve RR, companies enhance the product, simplify its usage, launch marketing campaigns targeting existing customers, and maintain constant communication with them.



Retention rate measures customer retention over time, aiding in evaluating product usefulness, reducing marketing costs, and boosting sales.

retention-rate-hoe-to-calculate.png

About what it means, why the Retention rate is used, as well as how to calculate the customer retention rate and increase the RR.

Retention Rate: How to Measure and Improve the Metric

Retention Rate: How to Calculate and Improve the Metric

<Video
url="https://www.youtube.com/embed/rAupz2rJSlQ?si=a9UHisSz1YvS33LD"
/>

In today's rapidly changing business world, it's crucial for businesses to look ahead and plan for the future. This is necessary to maintain stability, plan resource utilization, and develop growth strategies. Nowadays, companies turn to technology for "predicting the future," specifically to predictive analytics tools.

At the core of the method lies the use of big data and complex IT systems. Predictive analytics is essential for decision-making leading to increased operational efficiency, sales, and other business improvements.

In 2023, [**the predictive analytics market reached $14.71 billion**](https://www.fortunebusinessinsights.com/predictive-analytics-market-105179), and by 2032, this figure is projected to grow to $95 billion. Software for predicting future events based on data is being implemented in various business sectors and is becoming part of our daily lives.

<Img
img="/uploads/forecast_analytics_market_growth_6bd6b80d78.png"
alt="Popular methods of predictive analytics"
/>

In the article, we'll discuss what predictive analytics is, its benefits, and how this information analysis direction works.

## What is predictive analytics

Predictive analytics is the creation of forecasts of future events in business through working with existing (historical) data using statistical, mathematical, and [**machine learning**](https://altcraft.com/blog/methods-and-techniques-of-machine-learning) methods. As a result, the company obtains estimates of probabilities that may occur and impact the business.

Forecasting is a branch of data analytics. There are also descriptive, diagnostic, and prescriptive types of working with information.

Predictive analytics began to take shape in the 1940s during World War II. British mathematician Alan Turing and his team attempted to crack the German Enigma cipher machine. The algorithm changed daily, complicating the process, and it was necessary to "predict" the principles behind these changes. In the end, Turing managed to guess the commonality in the new combinations. It turned out to be a recurring phrase, through which they deciphered the principle of operation.

Later, IT technologies were incorporated into predictive analytics. Initially, the field where such algorithms were used was limited to financial departments. However, with the development of tools and changes in company needs, Big Data analytics became available to other business areas and employee categories. It began to play a role in decision-making for line managers and top executives. Now, not only historical data but also real-time information is used. This increases the efficiency and speed of forecasts.

## Methods of predictive analytics

Popular methods of predictive analytics include:

**Regression** — assessing the relationship between variables and the influence of one on another. It is used when processing large data sets. For example, this method determines how changes in price will affect sales.

**Decision trees** — classifying data based on specified variables. It is applied to predict what decision a person will make. The name is explained by the type of model, which resembles a tree. A branch represents a possible choice, and the leaves indicate the results.

**Neural networks** — a machine learning-based model used for tasks involving complex interactions. Neural networks are useful in situations where there are no specific formulas for processing data.

<Img
img="/uploads/popular_methods_of_predictive_analytics_107495b905.png"
alt="Popular methods of predictive analytics"
/>

## Benefits of predictive analytics for business

**Making objective and accurate decisions based on data** instead of relying on the experience of managers and guesswork. This reduces risks and avoids unnecessary resource expenditures on incorrect actions. For example, predictive analytics is needed for dynamic pricing.

**Maintaining competitiveness in the market.** The more information you obtain about future events, the better prepared you will be for changes. Then your offerings will be more precise than those of competitors.

**Increasing profits.** Forecasting indicates changes in demand, delivery times for goods, and can even predict when equipment essential for business operations will fail. Knowing the "future" helps the company adapt to the situation, hedge risks, and continue operating efficiently. Thus, the business will maintain and even increase its income by increasing LTV (customer lifetime value), as forecasting contributes to retaining customers longer through targeted marketing communications.


## What industries use predictive analytics

<Table>
| Sphere | How it is applied | 
| ------ | ------ |
| Sales | Estimation of supply and demand for up-to-date stocking of goods. Competitive pricing is also established and options for cross-selling are compiled based on customer preferences (user data is used for this). Retail predicts changes in demand based on seasonal and even climatic factors. | 
| Finance | In banking, risk predictions are important for loan disbursement. Machine learning algorithms are used to analyze the probability of whether a customer will be able to repay a loan or not. Predictive analytics also helps fight fraud in the financial sector. | 
| Marketing | Creating offers that hit the exact needs of customers. Precise advertising campaigns are created on the basis of the data, i.e. personalization through different communication channels works. This improves the quality of service, because people are not offered unnecessary content and promotions. | 
| Manufacturing | Predicting the timing of wear and tear and equipment failure in order to make repairs in time. It also predicts the volume of output, what factors may affect the output of the product. In addition, forecasting is important for production safety and accident prevention. | 
| Medicine |Patient history and observation data are used to develop disease variants and prevent them. Forecasting also works in organizing the work of clinics: on the basis of the data, the necessary amount of drugs is ordered and an optimal schedule is formed. | 
| Personnel Management | Predictive analytics "predicts" the compliance of employees with the company's values, calculates the possibility of employee dismissal and their reasons. Based on qualification data, lists of skills required for effective work are generated. |
| Logistics | Forecasting is important to build up sufficient stock levels in warehouses and to assess situations for deliveries at the right time. | 
| Cybersecurity | Potential methods of information theft are identified. This is how companies can minimize them or protect data storage completely. Cyberattacks are [**an urgent problem**](https://www.comnews.ru/content/231183/2024-01-29/2024-w05/1010/kolichestvo-kiberatak-2023-g-vyroslo-11) that companies face. Therefore, prediction in this matter is very important. |
</Table>

## Risks in predictive analytics

Despite significant advantages, predictive analytics is not a panacea.

- Firstly, there is always a risk that even the most advanced algorithm will make mistakes. Therefore, it is not advisable to fully rely on forecasts. <p>
</p>
- Secondly, predictive analytics may not consider all factors if there is insufficient data for analysis or if the data turns out to be inaccurate. <p>
</p>
- Thirdly, the emergence of a "black swan" event in the future is not excluded, rendering all forecasts unreliable. This is an unexpected and previously unknown factor. For example, the COVID-19 pandemic in 2020 turned out to be such an event for many business sectors, necessitating urgent process adjustments.

Predictive analytics streamlines business operations, but currently, this method cannot fully bear the responsibility for business decisions.

## How predictive analytics works

Predictive analysis relies on large volumes of data. Machine learning algorithms and complex algorithms are applied to process the information.

**The stages of data analysis for forecasting include:**

1. **Goal definition:** A specific business task is set that needs to be solved with data-driven predictions. For example, "determine the demand for product N in the future year." <p>
</p>
2. **Selection of forecasting methods** that will lead to the goal. It is important to identify the correct sources and types of data. <p>
</p>
3. **Data collection:** Information from various sources is integrated into a single system (e.g., [**CDP**](https://altcraft.com/en)). The specific data to collect depends on the sphere of analysis and project goals. For example, marketing channels provide information about user actions on the website, completed conversions, gender data, and so on. It is important that large volumes of data are required for accurate forecasting. <p>
</p>
4. **Information processing**, including data cleansing. At this stage, unreliable and incomplete data are removed. This is usually a lengthy but crucial process that affects the accuracy of forecasts. <p>
</p>
5. **Model building and validation:** Analysis parameters are set, based on which artificial intelligence provides forecasts. Model parameters may be adjusted as needed. <p>
</p>
6. **Result retrieval and implementation:** Based on the forecasts, customers (who may be managers) make strategic and other decisions.

## Summary

Predictive analytics is a data-driven approach where future events are forecasted based on past information. This method improves business metrics as managers make informed decisions, leading to increased profitability and market competitiveness.

Machine learning and methods such as regression, decision tree, and neural networks are utilized in predictive analytics. This approach is applied across various sectors including sales, marketing, finance, manufacturing, healthcare, human resources management, and logistics.

To make forecasts, goals are set, data is collected and processed, and predictive models are built. As a result, clients receive scenarios of future developments and make business decisions accordingly.


Predictive analytics forecasts future events from past data, enhancing business metrics and decision-making for higher profitability and market competitiveness.

predictive-analytics.png

About what predictive analytics is, what benefits it has and how this area of information analysis works.

Predictive Analytics: Definition, Where Used & How Works

Predictive Analytics: What It Is and How It Works

<Img
img="/uploads/what_is_statistical_power_a8256e4185.png"
alt="What is statistical power"
/>

Defining statistical power or the "sensitivity" of a test is an important part of planning before launching A/B testing. It allows you to introduce more positive changes to a website and increase your revenue.

## Definition of statistical power

**Statistical power** is the ability of a criterion to detect differences where they actually exist. The sensitivity of a test will show a significant result and reject the incorrect null hypothesis.

Before delving into the elements of statistical power, it's essential to understand the types of errors that can occur in tests and how to prevent them.

## Two types of errors

### Type I error

**Type I error** is a false positive result: rejecting the null hypothesis that is actually true.

**The null hypothesis** states that there is no difference between a pair of events or phenomena.

A false positive test shows that there is a difference between the compared variants, even though there isn't one in reality. Such a situation can occur due to an error in the test or if randomness creeps in.

The probability of a type I error is denoted by the Greek letter alpha (α) and indicates the threshold level at which an A/B test should be trusted. If a test is reliable at 95%, the remaining 5% represents the probability of a type I error.

If a business cannot tolerate a 5% chance of error, considering it too high, the probability of false positive results can be reduced by increasing the test's reliability. If the reliability is increased to 99%, the probability of a type I error decreases from 5% to 1%. However, there are some "pitfalls" in this calculation.

The decrease in the probability of type I errors increases the probability of type II errors because there is an inversely proportional correlation between these values. Therefore, reducing the alpha error level (for example, from 5% to 1%) ultimately reduces the overall statistical power of the test.

<Img
img="/uploads/types_error_36faa5eac7.png"
alt="Graph by statistical power"
src="Type I error and its proportional connection with type II error"
/>


Sometimes it's necessary to consciously increase the risk of type I errors (for example, from 5% to 10%) to achieve high test sensitivity.
### Type II error


**Type II error** is a false negative result, meaning the inability to reject a false null hypothesis. Such a test will not show the advantage of any of the options, even though there are options that are more effective than others.

The probability of a type II error (β-error) is inversely proportional to statistical power (1 - β). If the risk of a Type II error is 20%, the sensitivity of the test is 80%.

Since alpha and beta errors have an inversely proportional relationship, if a test has an extremely low alpha error value (for example, 0.001%), the risk of a type II error is very high.

Finding a balance is not an easy task. Through A/B testing, a business aims to determine the best combination of elements and identify the most effective variant among many. If the tests lack sufficient sensitivity, there's a great risk of not noticing a genuinely good option.


## Variables affecting test statistical power


Next, we'll examine four variables that influence the statistical power of a test. While studying these variables, it's important to remember that the primary task of a tester is to control the probability of errors. This can be achieved through:

- sample size; <p>
</p>
- minimum detectable effect (MDE); <p>
</p>
- significance level (α); <p>
</p>
- desired level of power (implied Type II error coefficient).


### Sample size


The sample needs to be sufficiently large; otherwise, conducting a good split test is impossible. It's crucial to calculate the sample size to ensure it's of the necessary and adequate size. A small sample won't provide enough test power, while an excessively large one will extend the test duration (long tests are more expensive and time-consuming).

Each analyzed variant or [**segment**](https://altcraft.com/help-center/manage-customer-database/database-segmentation-for-email-newslettering) should have a significant number of users. For this reason, the sample size should be planned in advance, ensuring that the test always has good statistical power. Otherwise, exiting the test might reveal too many variants or segments. If this realization comes too late, the test will have been conducted in vain.

Expect statistically significant results within a reasonable period (e.g., no less than a week or a business cycle). In most cases, it's recommended to conduct testing for 2-4 weeks. Any duration exceeding this might lead to sample contamination and issues with cookie file deletion.

Therefore, determining the sample size and testing duration beforehand is crucial. This helps avoid the common mistake of conducting A/B testing "blindly," ending before obtaining statistically significant results.


### Minimum detectable effect (MDE)


MDE is the difference between variants that needs confirmation. Small differences are hard to detect, requiring very large samples. Substantial differences are noticeable even in small samples.

However, there exists a paradox where small samples can be unreliable, leading to erroneous tests. There's always a risk of error, and there's a narrow range within which results can be considered reliable. Since there's no single rule for determining sample size, all nominal levels are not 100% reliable.

### Significance level


Testing results are considered statistically significant under the condition that the null hypothesis is incorrect.

In simple terms, if during a split test, one landing page can confidently be preferred over another (with a test confidence level of 95%), there's only a 5% chance that the observed improvement is due to chance or error. Conversely, there's a 95% chance that the observed difference is not due to chance.

The 5% significance level is the most commonly used initial level in online testing, representing a 5% chance of a type I error, as mentioned earlier. An alpha of 5% means there's a 5% chance of incorrectly rejecting the null hypothesis.

Under equal conditions, reducing alpha from 5% to 1% simultaneously increases the likelihood of a type II error. An increased risk of a type II error reduces the validity of the test.

### Desired level of power


If the statistical power of a test is 80%, there's a 20% chance that the real difference between variants won't be detected. If a 20% chance is too risky, it can be reduced to 10%, 5%, or 1% to achieve powers of 90%, 95%, or 99%, respectively.

Before concluding that a 95% or 99% power analysis will solve all your problems, note that each increase in analysis power significantly increases sample size and analysis time.

So, what level of test sensitivity do you really need? The acceptable level of false negative risk for conversion optimization is often considered to be 20%, corresponding to an analysis power level of 80%.

There isn't a universally recognized standard for 80% power, but it represents a reasonable balance between alpha and beta errors. Consider the following:

- What is an acceptable level of risk for missing out on a quality improvement for you? <p>
</p>
- What is the minimum sample size required for the test to achieve the desired sensitivity?


## Calculating the statistical power of a test

You can use an A/B test calculator. Enter known variable values to determine, for example, the required sample size to achieve sufficient test sensitivity.

For instance, you calculated that testing each variant requires a sample size of 681 individuals. Your calculations were based on the following initial values: test power of 80%, alpha of 5% (95% statistical significance). You also knew the conversion rate for the control group is 14%, and you assumed an expected conversion rate for the better variant of 19%.

<Img
img="/uploads/calculate_the_statistical_power_of_a_test_848e9bae89.png"
alt="Calculation of the sample size"
src="How to calculate the statistical power of a test"
/>


Using the same initial data, knowing the sample size, alpha value, and the desired level of statistical power (for example, 80%), you can determine the  minimum detectable effect (MDE).

<Img
img="/uploads/determination_of_mde_f4e64ce91b.png"
alt="Calculation of the minimum detectable effect (MDE)"
src="Determination of MDE"
/>


As we can see, an A/B test calculator is quite convenient when you know the values of three variables and aim to find the fourth one.

## What to do if the sample size cannot be increased?

In some cases, there's a need for a more high-powered test, but increasing the sample size is not possible, such as when the page traffic is too low.

Let's consider a scenario: you've inputted your data into the A/B test calculator, and it shows that a sample size of 8000 or more is required.

<Img
img="/uploads/use_a_sample_size_calculator_34fab3e500.png"
alt="Determining the required sample size"
src="How to use a sample size calculator"
/>

But you don't have the ability to gather such a sample; it would take several months or even longer. In this case, you need to increase the minimum detectable effect (MDE). In the example below, increasing the MDE from 10% to 25% reduces the sample size to 1356.

<Img
img="/uploads/sample_size_depends_on_increasing_the_mde_1aac156651.png"
alt="Increase the MDE to get the desired sample size"
src="Sample size depends on increasing the MDE"
/>


Is it really always possible to increase the MDE to 25%? After all, this will reduce the quality of the test.

The second option comes into play if you're willing to accept a 10% risk of type I error. In that case, the optimal option would be to decrease the test's significance level to 90%.

<Img
img="/uploads/balance_between_errors_within_calculator_1d200f99e4.png"
alt="Reducing the level of confidence to obtain the required sample size"
src="Balance between errors within sample size calculator"
/>

Iterating through data in hopes of obtaining satisfactory figures isn't the most reliable strategy. Analysts recommend starting by determining the sample size and then arbitrarily adjusting other values until the results are deemed satisfactory.

## Conclusion

Statistical power helps avoid errors during testing and identifies genuinely impactful performance factors on your platforms.

Adhere to these simple guidelines:

- Choose the test duration wisely: ideally 2-4 weeks <p>
</p>
- Use a test calculator to determine an acceptable test power <p>
</p>
- Adhere to the requirements for the minimum sample size <p>
</p>
- Aggregate segments and test them if needed <p>
</p>
- Set sensitivity test requirements only after meeting the minimum sample size requirements.

Source: [**CXL**](https://cxl.com/blog/statistical-power/)


Statistical power gauges a test's ability to detect differences. It helps avoid false conclusions by assessing the test's sensitivity to find genuine changes.

what-is-statistical-power.png

About how to correctly calculate the statistical power for testing.

Statistical Power: Definition, How to Calculate & Variables 

What is Statistical Power

<Img
img="/uploads/audit_dannyx_014ad471e9.png"
alt="How to Maximize Your Buisness Decisions with Data Auditing"
/>

The world generates [**2.5 quintillion bytes**](https://explodingtopics.com/blog/big-data-stats) of information every day, with the Big Data market reaching $274 billion in 2022 and the analytics platform market following suit. This upward trend shows no signs of slowing down.

Data is highly valuable for businesses, serving as an inexhaustible source of inspiration for marketers and [**providing insights**](https://martech.org/analytics-audit-101-identify-issues-and-correct-them-to-ensure-the-integrity-of-your-data/) into consumer habits and expectations. Through data, companies can make informed predictions about their growth and select effective methods for outcompeting rivals.

However, processing large amounts of information is a challenge. Without quality analytics, the data collected is nothing more than meaningless numbers.

Data analysis, first and foremost, involves checking the authenticity of the information. But that's not all. The main task of the audit is to determine how the collected data can help the company solve the set tasks. Without accurate data, companies can't make informed decisions. Let's explore data auditing in more detail.

## How data auditing can benefit your business

Data can be a lifesaver for a company, as long as the sheer amount of it doesn't overwhelm management. Storing a large volume of information can be costly, and it needs to be organized and utilized correctly. This is where a data audit comes in. Data auditing can:

- **Consolidate disconnected databases from various departments.** Auditing, using cloud tools, can combine information across the entire company. This allows marketers to utilize valuable information collected by the sales department.
- **Resolve data integrity issues.** Company databases often contain duplicated, incomplete, or incorrect information, which cannot be used as a basis for business decisions. Relevant and reliable data is needed to solve business problems.
- **Comply with regulations.** To maintain order in business processes, many companies establish internal regulations for handling information, while some industries are regulated externally. Auditing helps ensure compliance with prescribed rules, avoiding fines for violations of data storage and usage deadlines.
- **Avoid data leaks.** [**Data leaks**](how-to-respond-to-data-breaches) occur when third parties gain unauthorized access to information that is not intended for them. Personal data security issues can be significant for companies, which is why they are vigilant in this area. Data auditing highlights vulnerable areas.

## Phases of data auditing

### Step 1. Planning

At this stage, it is necessary to develop a step-by-step plan for the audit and assign responsible individuals for its preparation, conduct, and presentation of results. The success of the entire «operation» depends on how well the planning is done.

### Step 2. Inspection and information gathering

Find out what information is currently stored in the company. How it is collected, how long it should be stored, and most importantly, how it is used in business processes.

### Step 3. Analysis and evaluation of data collected

Evaluate the situation for compliance with legislation and regulatory acts. Check data security and determine the procedure for employees' access to information.

### Step 4. Development of recommendations

Develop correct business processes and implement them in practice. Reject and replace non-working algorithms. Provide recommendations to employees on how to work with information. Remove excess data, set current storage periods for information that quickly becomes obsolete. Provide access to information for those employees who need it for their work.

<Img
img="/uploads/data_audit_steps_7cd0afe42d.png"
alt="Phases of data auditing"
/> 

## Who is in charge of data auditing?

Data scientists are usually the ones responsible for conducting data audits. In some cases, it may be appropriate to form a team of competent individuals to monitor information usage, make budget decisions, and set priorities. The team's work typically results in internal policies and data processing strategies that are then shared throughout the organization. This is how employees are trained to work with data within the company.

## Why is data audit automation needed?

Based on the analysis of events of the past three years, it can be confidently stated that customer information quickly becomes outdated: their behavior and desires change so rapidly and unexpectedly. Data quality auditing is necessary for modern businesses, as it provides the opportunity to make relatively accurate predictions of the near future. However, processing large amounts of information manually is practically impossible.

This is where data management automation platforms come in. They completely take over routine and labor-intensive processes such as cleaning the database of duplicates, monitoring information sources, and clearing the database of clutter. All of this is necessary, but it is not practical to do it manually.

Automating data auditing is beneficial for the company. It significantly frees up employees' time that they can spend on more critical and creative tasks. Moreover, the results obtained are more accurate and reliable than if the database were processed by a human. Robots do not make mistakes, do not get distracted, do not tire, and their vision does not «blur». Therefore, in this regard, they work much better than a human.

## How to choose a data audit automation tool

Platforms designed specifically for collecting and storing marketing data can help you with this task. Artificial intelligence has made significant advancements in this area, enabling precise targeting of marketing campaigns and reaching interested audiences.

Smart platforms identify new visitors using JavaScript tags and track their IDs, gradually building detailed profiles that include behavioral characteristics, demographic data, and purchase history. The data is not mixed up between users, processed accurately, and subjected to analysis.

When selecting a platform, it is important to look for the ability to [**segment users into target audiences**](how-and-why-to-segment-target-audience) in real-time. Businesses find it easier to succeed in the market when they have accumulated enough useful information. The key is to ensure that this information is relevant to improving current business processes within the company.

In this post , we will explain what a data quality audit is, how it can help your marketing, and how to conduct a data audit.

audit dannyx.png

For those who want to know what a data audit is and how to conduct it.

How to Maximize Your Buisness Decisions with Data Auditing

Data Quality Audit: How It Can Benefit Your Marketing

<Img
img="/uploads/ROAS_ROI_ROMI_7292d14939.png"
alt="ROAS, ROI, ROMI"
/>

If your company only invests money in development, but does not earn it, that's definitely something to think about. ROAS, ROI, ROMI metrics are used to evaluate the effectiveness of investments in business and individual areas of work and, as a result, to see the weak spots and make the right decisions to change the strategy.

Each metric is suitable for a different degree of company overview. ROI is a complete assessment of the success of investments; it takes into account all revenues and expenses. ROMI is calculated when working with marketing. ROAS is used for individual advertising campaigns.

Metrics are often confused, so in this post we are going to figure out what they mean, how they differ, and why each is needed. We will also tell you how to calculate ROAS, ROI and ROMI correctly.

## What is ROI

**Return on Investment**, or ROI is used to understand how much is earned from each dollar invested. The ROI formula takes into account the costs of creating, maintaining, and promoting a product. This includes the cost of purchasing materials, salaries, the budget for advertising campaigns, etc.

### Why and how to calculate ROI

The calculation of ROI is important for assessing the profitability of your business: whether investments pay off and whether there is a profit from them. When the ROI is low, the decision is made to close unprofitable projects or change the marketing strategy and the entire direction of the company's development. ROI saves you from wasting your budget and tells where to invest money to get a better result.

**ROI formula (%):**

<Img
img="/uploads/ROI_b6b5495f5a.png"
alt="ROI formula"
/> 

ROI is calculated for a certain period, but during this period there should not be a dramatic change in circumstances, such as price hikes, new suppliers, and so on. In the formula, **revenue** is the money earned by the company. **Costs** include all investments made.

Also, the time interval for calculation depends on the sales cycle length.

What is a good ROI? Usually a result above 0% means that you have made a profit. Below that figure means the business is losing money. Exactly 0% ROI means that the business has paid off the investment, but has not earned anything.

*Example. In September, a sportswear store spent 10,000 dollars to update its clothing collection, paid employees 5,000 dollars in salary, and bought advertising for 3,000 dollars. This month, the store's sales amounted to 22,000 dollars. ROI = (22000-18000)/18000=22%. This is not a high value, but at the same time, the costs paid off in full, and the store's activities in September made a profit.*

### How to increase ROI

* **Check the effectiveness of marketing channels** and disable or adjust those that bring low results.
* **Work with your target audience.** It is possible that [**low quality leads**](MQL-and-SQL) are coming in because you are addressing the wrong people in your ads.
* **Connect sales and marketing** if there is inconsistency or problems getting information to another department. Sometimes a successful advertising campaign comes to naught if the sales team does not handle leads the right way.
* **Revise your business goals** and make sure they are realistic. Exaggerated plans will not bring results if there are not enough resources.

## What is ROMI

**Return on Marketing Investment**, or ROI is used to understand how much is earned from every dollar invested in marketing. Marketing expenses include the costs of advertising, promotions, events and other marketing activities. It can also be calculated for individual channels, such as [**messenger**](messenger-marketing) or email campaigns, contextual advertising, and so on.

### Why and how to calculate ROMI

The metric is important for evaluating the effectiveness of promotion. ROMI gives businesses a broader view of marketing. If you evaluate the superficial result (the number of clicks, applications, etc.), you can decide that the campaigns are successful. However, after calculating the return on investment, it turns out that you spend more money on marketing than you earn from it.

**ROMI formula:**

<Img
img="/uploads/ROMI_0dbda34671.png"
alt="ROMI formula"
/>

The ROMI formula is similar to the ROI calculation, but the **costs** include only the costs of marketing tools.

*Example. The company spent 3,100 dollars on contextual advertising, 1,000 dollars on email campaigns, and 900 dollars on [**influencers to promote products**](how-to-find-effective-influencers-to-promote-small-business). In the current month, the company's income amounted to 4,500 dollars. ROMI is (4500-5000)/5000= –10%.*

**What is a good ROMI?** More than 0% means that the investment pays off. Less than 0% — costs exceed revenue. 0% shows that investments go «to zero». *In the example above, the investment does not recoup the costs.*

### How to increase ROMI

* Make it easier for users to buy. Remove unnecessary steps in advertising and on the site. Don't force interested leads to fill out dozens of fields and confirmations. Let them reach the “Buy” button faster.
* Check the effectiveness of channels and individual advertising campaigns. Don't invest in things that don't work. Always look for ways to optimize your advertising and website — test hypotheses and reject unsuccessful solutions. Invest more in high-priority channels that generate revenue.

<Note>
Reliable data management system increases control over marketing performance. [**Altcraft Platform**](/en) automates the work with advertising and collects information from different sources in a single profile, so the company knows everything about customers. As a result, marketers make the right decisions and choose profitable strategies.
</Note>

## What is ROAS

**Return on Ad Spend**, or ROAS is a measure of the return on investment a business spends on advertising. It is used to check how much money has returned from investments in individual ad campaigns or channels. The ROAS formula includes advertising revenue (due to customers who saw the ad, clicked on the link and bought) and campaign costs (usually the cost of advertising on sites and excluding the work of a specialist).

### Why and how to calculate ROAS

ROAS is a metric that shows the effectiveness of a particular advertising campaign or promotion channel. You can evaluate individual strategies and techniques to eliminate unsuccessful ones and allocate a larger budget for successful ones. Thus, ROAS is needed to make decisions on specific marketing actions, but not on the entire strategy, as ROMI.

**ROAS formula:**

<Img
img="/uploads/ROAS_00f0e1c5ce.png"
alt="ROAS formula"
/>

The total metric for a campaign or individual ads is calculated.

*Example. The online store launched a Facebook advertising campaign and sent out an email campaign. Facebook advertising expenses amounted to 1,000 dollars, and the revenue received was 1,800 dollars. The email campaigns cost the company 300 dollars, and the revenue was 1,200 dollars. The ROAS of an advertising campaign on Facebook is 1800/1000=180%. And the ROAS of email campaigns is 1200/300=400%. Perhaps the company should increase the investment in email marketing, since it brings more profit for every dollar invested in advertising.*

**What is a good ROAS?** The calculation is different from ROI and ROMI. ROAS is considered good if its value is above 100%. Otherwise, investments in advertising do not pay off.

### **How to increase ROAS**

* Change creatives and offers of advertising campaigns that are not performing well. Check them out through testing.
* Drop the channels that bring insignificant results.
* Find weaknesses in the [**sales funnel**](what-is-sales-funnel) and work out the places where leads are lost.
* Check the effectiveness of the site: clarity, speed, accessibility, etc.

## The difference between ROI, ROMI and ROAS

The metrics are similar, but differ in scale. Some show the business globally, others in more detail. Let's compare them in the table.

<Table>
| **Difference** | **ROI** | **ROMI** | **ROAS** |
| ------ | ------ | ------ | ------ |
| **Calculate the return on investment in…** | business | marketing | ad campaigns |
| **Used to…** | Сhange the overall business strategy and focus areas | Test effectiveness and adjust marketing strategies | Change or cancel advertising campaigns |
| **What costs to consider…** | all (creation, maintenance, product promotion, and so on) | only marketing | only the budget of individual campaigns |
| **Formula** | (Revenue − Costs) / Costs * 100% | (Revenue − Marketing costs) / Marketing costs * 100% | Campaign revenue / Campaign cost * 100% |
</Table>

## Key points when working with ROI, ROMI and ROAS

* **Consider the context of the metrics.** Pay attention to the circumstances (market conditions) and don't look at metrics in isolation.
* **Collect accurate and complete data for the right analytics.** Choose a convenient system for data management.
* **Correctly determine the calculation period**, which depends on the duration of the sale. The longer the deal, the longer the time span.

## Conclusion

ROAS, ROI, ROMI are metrics that help to understand the effectiveness of investments in a business: whether the money spent on its development pays off. ROI refers to more global and general metrics. It takes into account the return on investment after calculating all investments made (from the cost of producing to advertising). ROMI shows the effectiveness of marketing investments. ROAS is responsible for the return on investment of individual advertising campaigns.

Metrics give a hint what to change in the work of the company. ROI is needed to adjust overall strategies. ROMI is calculated to adjust marketing strategies. ROAS answers the question of what to do with specific ads, newsletters, and other campaigns.

It is important to consider all metrics in context with other business data and use analytics systems.

Metrics are often confused. Read the post to learn what these metrics mean, how they differ, and how to calculate ROAS, ROI and ROMI correctly.

ROAS, ROI, ROMI.png

For those who do not know how to calculate the return on investment in their business.

ROI, ROMI, ROAS: What is the Difference and How to Calculate

What is the Difference Between ROI, ROMI and ROAS

<Img 
img="/uploads/Blog_b0169471f4.png"
alt="Sales funnel"
/>

Customers do not always buy as soon as they find a product on a company's website or elsewhere. When it comes to complex and expensive products, consumers take the time to compare and study. For this reason, companies try not to be intrusive and help users go through all stages of decision making step by step.

It is a sales funnel that is used to make the process transparent and simplify analytics. A sales funnel is a concept that describes all the steps a customer takes on the way to a conversion. There are several funnel models: the classic five-step buying scheme, the AIDA model, the SPIN method, and others. The company chooses a model and adapts it to its needs or creates its own version.

In this post, we will explain what a sales funnel is, why use this concept and how to develop it.

## What is a sales funnel and why use it

A **sales funnel**, also known as a **marketing funnel**, is the journey a customer takes from first contact with a business to making a purchase. The model is usually visualized as an upside-down cone and represents all the stages of sales that the client goes through. The top part gets the most [**leads**](MQL-and-SQL). But not everyone gets to the end where the purchase takes place, so the cone narrows to the bottom.

### Why does your business need a sales funnel

- **Highlight weak spots.** When all steps are written and all stages of the sales funnel are tracked, it is easy to analyze and understand where there is a problem in sales.
- **Supervise the work of sales reps and marketers.** The funnel includes marketing tools and sales activities, so it is easy to track the effectiveness of each participant in the process.

*For example, your advertising campaign results in 100 applications per week, but only 5-10 people buy. The analysis shows that the applications are targeted, and most of all clients are “turned back” after talking with the manager. You look into the situation and find that sales reps are not interested in dealing with the client and do not know how to handle objections about the high price. If the company did not have a sales funnel, it would be more difficult to find the reason.*

- **Create controlled sales.** If you know at what stage of decision-making the lead is now, you can make an offer that will satisfy their needs. This approach will speed up the purchase.
- **Analyze sales performance.** You can calculate your sales funnel conversion using the formula. Divide the number of converted customers by the total number of users, and then multiply the result by 100%.

<Img
img="/uploads/sales_funnel_conversion_formula_db6b695bd1.png"
alt="Sales funnel conversion"
/>

In other words, this is the difference between the number of leads at the top and the bottom. The more potential customers reach the purchase, the better the funnel works.

## Sales funnel stages

The classic sales funnel is the AIDA model, which consists of four stages.

<Table>
| **Stage** | **Description** | 
| ------ | ------ |
| **Attention** | The lead contacts the brand for the first time through advertising or direct search on the Internet. |
| **Interest** | The lead is interested in the product, for example, they subscribe to the newsletter, click on the links or ask questions. |
| **Desire** | The lead understood that he needs the product, and now they are looking for options for a profitable offer, such as discounts, free shipping, etc. |
| **Action** | The lead places an order. |
</Table>

<Img
img="/uploads/classic_sales_funnel_8977a756a3.png"
alt="Classic sales funnel"
/>

The other type is the marketing funnel, which repeats the five stages of sales.

<Table>
| **Stage** | **Description** | 
| ------ | ------ |
| **Approaching the customer** | The first contact between a sales rep and a potential client or automatic first message. It is important to make the right impression and motivate a person to continue communication. |
| **Discovering customer needs** | The sales rep asks the client questions about their needs. Then the manager draws a conclusion and offers a suitable solution. |
| **Pitching** | Present your solution. Based on the discovered needs of the client, talk about the features of the product and its benefits. The sales rep presents only those features that are important to the client. |
| **Handling objections** | The sales rep answers the buyer's questions and works out objections. For example, doubts about the cost or quality of the product. |
| **Closing the deal** | The customer is ready to buy and is interested in payment methods or installments. The task of the sales rep is to switch the client to the completion of the purchase: pay, order delivery, and so on. |
</Table>

In brick-and-mortar retail stores, all five steps are done in a conversation here and now. However, if your company operates in the digital space, the online sales funnel consisting of such stages is not always fast. Sometimes the process includes phone calls, messaging, and sending files.

Another option is the SPIN method. Its author, Neil Rackham, said that classic five-step sales models are not effective for large and complex deals, such as B2B. According to SPIN, at each stage managers work with different types of questions that identify the customer's problem, clarify and lead to a purchase.

<Table>
| **Stage** | **Description** | 
| ------ | ------ |
| **Situation** | Open-ended questions to clarify the situation in which the customer is now.  |
| **Problem** | Questions to identify problems and hidden needs. |
| **Implication** | Questions that demonstrate the possible negative effect if the problem will not be solved. |
| **Need-payoff** | A type of “what if” question that guides the customer into the future; questions like these show how a customer's life would change if they fixed a problem with your product. |
</Table>

Each company chooses the model that best suits its needs, and transforms the stages, adjusting them to the business processes.

## How to build a sales funnel

There are no universal methods, so brands not only adapt existing marketing funnel models for themselves, but also develop their own. Some models also include after-sales service, upselling and cross-selling.

Let's create a custom sales funnel in six steps:

### 1. Collect data

Collect information about the audience that should be included in the list of leads. Determine the channels through which customers come. Find out how customers behave on different resources, as well as what their needs and objections are. Next, plan a funnel for a specific audience, and not for all customers at once. Choose one of the existing models, exclude the steps that are not important for your company and add the missing ones.

### 2. Define a goal for the funnel

Specify what conversion rate you would like to receive from the funnel and also rates for intermediate stages. The model is divided into top, middle and bottom. Specify how many users are expected to reach each stage.

### 3. Think through the steps of the funnel

Define actions at each funnel step. For example, at the beginning it may be the launch of an advertising campaign. The middle is a call from the sales rep or an email reminder of the purchase. At the end, you can offer free shipping or bonuses.

### 4. Create content for each step

For automatic funnels, the content can be email templates. If sales reps lead customers through the funnel, prepare scripts, materials that they will send to customers for a detailed acquaintance with the product.

### 5. Set up a sales funnel

Marketing funnel can work:

- manually through the sales staff;
- semi-automatically — email marketing services, phone services and other automation platforms are responsible for some of the stages;
- automatically — a marketer only controls and optimizes.

You will have to use [**special services**](/features/analytics) to set up semi-automated and automated sales funnels. In the future, this will save time for your employees and speed up communication with the client.

### 6. Run the funnel, analyze and optimize

Once the funnel is up and running, it is important to monitor and improve it. Regularly measure performance at each stage and quickly respond to changes. Only in this way the model will bring the expected income. To optimize your funnel, you need to get feedback from users on the sales process and analyze the stages of the funnel.

## Sales funnel example

The user visits your site, adds the item to the cart and immediately checks out. This is an example of a simple funnel, but this rarely happens. Often the path to purchase is longer and more difficult. Here is an example of such a funnel.

The user:

1. Gets to the company's website through contextual advertising. 
2. Views a product, for example, an autumn jacket. 
3. Subscribes to an email newsletter that informs about promotions and brand news. 
4. Receives an email with an offer for the autumn collection. 
5. Follows the link and adds the item to the cart, but does not check out. 
6. The company sends a reminder email about the product in the cart and the user buys.

## How to automate your sales funnel using Altcraft Platform

The platform supports [**omnichannel marketing**](omnichannel-vs-multichannel), so it collects information about different stages of the sales funnel in a single profile. Let's take a look at how to automate the customer journey from the example above.

1. You set up and run contextual advertising through Google Ads.
2. Jane visits your site and leaves her email. Her contact details are included in the Altcraft Platform database. The subscription to the resource is added.
3. The platform launches an automatic series of welcome emails.
4. Jane receives an email with a promotional code. She reacts and adds the item to the cart, but does not check out.
5. Altcraft Platform starts an [**abandoned cart**](how-to-reduce-cart-abandonment-rate) email scenario.
6. Jane becomes interested and buys. The funnel ends.

The entire process of moving a user through the sales funnel is automated through workflows and other Altcraft Platform tools. Sending messages through different channels — Email, SMS, Push, WhatsApp or Telegram — can expand the possibilities of your business and make the customer journey completely clear and controlled.

<Img
img="/uploads/abandoned_cart_scenario_607a42d15b.png"
alt="Abandoned cart scenario"
src="Abandoned cart scenario"
/>

The Funnel report is used to track the process performance. The report displays metrics for all actions within the strategy. It is possible to visualize all the steps of the sales funnel: the number of customers and the percentage of conversion.

<Img
img="/uploads/sales_funnel_in_altcraft_platform_d15cdd4757.png"
alt="Conversion funnel in Altcraft Platform"
src="Conversion funnel in Altcraft Platform"
/>

## Conclusion

A sales funnel is a model of a customer's journey from the first meeting with a brand to purchase. Using this approach, you will make sales transparent and controlled. The company begins to control each stage of the transaction, monitor the work of the sales department and find weak spots.

Building a sales funnel is possible using the AIDA, SPIN models, five classic sales steps and others. There is no universal solution, so companies also create their own models. To do this, you need to collect customer data, define goals, make a plan for each stage, prepare content, set up a funnel, and continue to optimize after launch.

This post explains what a sales funnel is, why use the concept, what funnel models exist, and how to build your own marketing funnel.

Блог.png

A sales funnel is a concept that describes all the steps a customer takes on the way to a conversion.

Why Businesses Use a Sales Funnel and How to Build It

Sales Funnel: Definition, Models and Stages

<Img
img="/uploads/Blog_1_320fd9f520.png"
alt="Why You Need Data Modeling"
/>

Business cannot exist without data. Companies plan, analyze and solve their problems using data. Information is collected from everywhere at each stage of the organization's activities. Modeling is needed to make the most of the collected data.

## What is data modeling

Data modeling is a method of compiling and structuring information. You don't need special technical skills to do data modeling. All you need is logic and understanding of business processes.

## What are the benefits of modeling

Here are some of the challenges data modeling can help you solve in business:

- Organizes business processes;
- Provides storage, systematization of information and make it easier to find the information you need;
- Sorts extra data and fills in gaps.

Proper organization of data ultimately leads to effective management decisions.

## How to model data

Start by building a conceptual model, then create a logical model based on it, and finally a physical one. 

### Conceptual data modeling

This is a visual representation of business processes and the relationship between them. This stage does not require software; it is enough to draw a model by hand.

### Logical data modeling

You create a conceptual model in the first step, and then you can convert it into a logical data model. To do this, add key and non-key attributes to the entities. And also you can show the relationships between primary and foreign keys. The result is an independent model that is compatible with all databases.

### Physical data modeling

The entities are transformed into tables; the attributes are transformed into columns. The names of the table and columns within it are adapted for compatibility with databases.

For ordinary users, the physical model is difficult to understand. Use the logical concept for dialogue with users, and use the physical one for in-depth analysis.

## Data Modeling Approaches

### Hierarchical data model

The simplest model was presented by IBM in 1996. In the approach, much attention is paid to the hierarchy based on the principle of a tree structure. The «parent» entity is located at the top, «children» goes down from it.

[**Information Management System**](https://www.ibm.com/ibm/history/ibm100/us/en/icons/ibmims/) is great for taking inventory. In a clothing store, at the top of the model is «clothing», its «children» can be «male» and «female». The decomposition separates women's wardrobe items (tops, dresses, skirts) from men's wardrobe items (trousers, shirts, jackets).

<Img
img="/uploads/hierarchical_data_model_a6083d81ed.png"
alt="Hierarchical data model"
/>

### Relational data model

This model was invented by Edgar Codd in 1970. Edgar, an IBM employee and computer scientist, based the model on an intuitive way to create tables. The data is organized in columns and rows, each cell is in its place. This method is still used effectively in the field of enterprise computing. Specialists use this method in the field of enterprise computing. Using relational models, it is convenient to track the number of goods in stock, store information about customers and conduct transactions in e-commerce.

### ER data models

This model stands for «entity-relationship». Within the model, entities are connected using relationships.

For example, when a user places an order in an online store, the order is assigned a unique code. This is an entity. Further, many events can happen with this order: the client can add or remove items, change the quantity, or even make a return. All these actions are the relationships to the entity.

### OODM

The Object Oriented Data Model (OODM) maps any complex scenarios. A closer look highlights the obvious and not so obvious business problems. OODM has proven to be great for funnel analysis. A potential client acts as an object with its own attributes (name, email, phone, etc.). As long as the lead has not [**turned into a client**](MQL-and-SQL), there are two objects: a «potential client» and a «client» itself. The «client» has all the same attributes as the «potential client», but with additions: ID, addresses and payment details.

OODM is designed to build relationships around a real-world object. When specialists build such a model, they rely not only on logic and mathematics, but also on the actual situation.

### Dimensional data model

When there is too much information, it is divided into data of the first, second and subsequent orders. In such a structure, it is very difficult to quickly find the required solution. Dimensional models are used to optimize databases and make it easier to find the information you need.

<Img
img="/uploads/dimesion_data_model_fa8c1d26c3.png"
alt="Dimensional data model"
/>

It is recommended to use OLAP (online analytical processing) for trend analysis, sales forecasting, financial reporting, etc.

## How to build an efficient data model

1. **Set a clear, measurable goal.** The goal should work for the result. Imagine what effect you want to achieve through model building.
2. **Strive for simplicity.** This advice applies to both the initial stage of working with the model and subsequent ones. At the very beginning, do not try to make the model complex and confusing. Get maximum clarity. In the future, you will fill it with the necessary information without making it difficult to understand.
3. **Organize your data based on facts, dimensions, filters, and order.** Let's say a business owner has over a hundred stores. The task is to find the point that is the most effective in terms of sales of product N over the past year. "Fact" is the sum of sales of each specific store. "Dimensions" are the product N that the owner is interested in, and the location of the store. "Filter" sets the time interval (year), and "order" determines the Top 10 stores according to the specified parameters.
4. **Avoid mistakes.** This is especially important when it comes to a large-scale project. If you skip a mistake to the next stage, there is a chance that at the very end of the project you will have to return to the roots. To avoid such situations, double-check.

## Data modeling tools

There is a wide range of data modeling tools on the market. In this post, we list a few of them. When choosing a tool, rely on the needs of your own business and pay attention to the following points:

- How clear is the organization of work with large amounts of data?
- Is there a visualization functionality?
- Can this tool streamline data mapping through the enterprise?

Here are some options:

- [**Erwin Data Modeler**](https://www.erwin.com/products/erwin-data-modeler/) is a tool from Quest. Suitable for designing and visualizing data and compiling detailed reports. 
- [**Enterprise Architect**](https://sparxsystems.com/) does a great job of visualizing business processes. 
- [**ER/Studio**](https://www.idera.com/products/er-studio/enterprise-data-modeling) is a service for working with complex data; you can build dimensional and relational models in it.
- [**Altcraft Platform**](/en) is a marketing automation platform that combines data from various sources into a single profile. Analytics works in real time, visualization is achieved through charts and reports. The advantage of the tool is the speed of processing large amounts of information.

<Img
img="/uploads/summary_report_d793cf8f9d.png"
alt="Data modeling in Altcraft Platform"
src="Data modeling in Altcraft Platform"
/>



In this post, we talk about what data modeling is, why it is important, and what tools to choose for data modeling.

Блог (1).png

Most of your data will come from your business processes and software system. Modeling is used to maximize value from the data.

Data Modeling: What It Is and What Tools to Choose

Why You Need Data Modeling

uzum-case.png

Reduced push notification opt-out rates, increased OR, saved working hours

bnb-bank-kejs-v-menyu.png

The open rate for automatic Viber messages is around 60%, and MAU growth for the year is +19%

All user information is systematically collected in a unified customer data platform

Distributing messages across various communication channels: email, SMS, push notifications, messengers

Promo codes, bonuses, and points for building long-term customer relationships

You can create and edit any HTML template, as well as save your own custom blocks

Personalized distribution of messages to clients through communication channels

Conducting experiments and optimizing marketing strategies

Tracking the effectiveness of marketing campaigns to make informed decisions

Connecting external systems to the platform and synchronizing data with company databases

Creating questionnaires, surveys, and feedback forms of any level of complexity

The ability to communicate with your clients anytime and anywhere

Increasing user awareness of your company's offers

Customer Data Platform

Omnichannel Marketing

Loyalty programs

HTML Email Builder

Personalized mailings

A/B/n Testing

Real-time analytics

Integration and WebHooks

Forms and Surveys

Messenger Communication

Pop-Up Windows

Interesting and useful articles about marketing, our product and online communications

Don't forget to subscribe to our Blog