Data retention periods and their impact in analyzing user behavior

Data retention periods and their impact in analyzing user behavior

Renata Ekine
Renata Ekine (Smartlook Team)  |  Last updated: Mar 6, 2023
10 mins read
Successful user behavior analysis is impacted by the data retention period; the period for which you store your data.

Google Analytics and most other analytics solutions track, measure, and report user actions regarding websites and mobile apps in terms of numbers (quantitative analysis).

On the other hand, advanced behavior analytics tools like Smartlook (in addition to the capabilities above) help you understand the why behind user actions — why users converted (or didn’t). Do users experience frustration while navigating your website or app? What about the overall user experience?

In both cases, either set of analytics data doesn’t allow you to derive much actionable insight when accessed only for a short time.

In this article, we’ll discuss:

Data retention and data retention periods

In analytics tools, data retention (aka data storage) refers to the period in which analytics data is stored. To put it simply, you only have access to the data for a limited amount of time.

Data retention periods vary across analytics tools. Typically, the minimum data retention period is roughly 30 days, with the maximum being “always/never expires.” Although there isn’t one data storage period that suits all businesses, the longer the data retention timeframe, the better. 

The benefits of extended analytics data retention periods

When analyzing user behavior, whether on your website or mobile app, the more data you have access to, the better you can spot patterns and trends, detect errors, and provide an enhanced user experience. 

In addition, you can segment your data into smaller, more digestible chunks without sacrificing the accuracy of your insight.

Throughout this article, I will be referring to real-life use cases to highlight how having access to historical data can help you better understand user behavior and gain insight into what’s working and what’s not concerning your website or mobile app.

Try Smartlook for free

Get a free 30-day trial, with all the features from the Enterprise package

Real use cases that highlight the importance of extended data retention periods

The ideal data retention period depends on your business goals. The following advice will help you understand what period is suitable for your needs. 

Continue reading to understand how access to historical data with a large lookback window can help you optimize the user experience. 

1. More analytics data = accurate insight

To make a meaningful analysis that translates insight into action, you need access to plenty of data. Having a broader timeframe of analytics data will allow you to understand the impact of marketing activities and seasonal promotions. Furthermore, you’ll be able to better assess the release of new app versions, features, and landing pages.

For example, when releasing a new homepage, it is important to compare its performance to its predecessor. 

Heatmaps are a great way to view aggregated data in a single image and see what users click on. With Smartlook, you can generate retroactive heatmaps in minutes using existing historical data rather than the data collected the moment you create a heatmap. 

To put it simply, if your tool’s data storage allows you to retain data for 3 months (as Smartlook does by default for Business and Custom plans), then you can build heatmaps based on your 3-months of existing data. This comes in handy when using heatmaps to visualize changes in user behavior. You can generate heat maps that represent various versions of your homepage monthly or at other points in time.

Take a close look at the screenshots below to visually reference how user behavior changed from February (top right) to March (bottom left). From this month-over-month comparison, we can conclude that running more remarketing campaigns that direct people to a homepage leads to users being more interested in pricing than features. 

2. Accurately measure the impact of your enhancements

One of our clients, StoragePug, relied on analytics data to evaluate one of their newly launched features. 

The StoragePug product team launched a scheduled email report feature as an alternative way to access essential data rather than visiting their Insights platform.

With the help of session recordings, they could zero in on users who experimented with their new feature. Insights from the session recordings helped validate their hypothesis that users prefer to access data via scheduled email reports rather than by visiting the Insights platform itself.

Based on these findings, after determining that their new feature remained stable, the team decided to promote it to their entire user base.

The ability to access analytics data related to scheduled email reports allows their product team to create a funnel to understand the benefits of feature adoption. In this case, having an extended data retention period will enable them to better understand how StoragePug users interact with their new feature. 

Keep in mind that launching a new feature isn’t something you “set and forget.” You need to constantly check your system to ensure everything is working correctly and that your users find value in it. Over a year or two, you may discover that a feature dropped from favor, from being used 2,000 times a month to 400. This is indicative of a serious problem.

In this case, having access to feature-related data dating back 12 months (or even more) can help you evaluate the long-term effect a new release has on the performance of your website or app. 

3. Detect errors and isolate bugs

Launching new features doesn’t always go smoothly. The product team at StoragPug found this out the hard way. They noticed that once they launched their latest feature, their users were unable to place orders. 

Typically, their team would wait for a day or two to collect enough quantitative data before declaring something amiss and launching an investigation. Still, this time, their PM didn’t want to wait. He opened Smartlook, looked over some recent session replays, and quickly noticed that their UI was not rendering correctly. He immediately notified their engineering team, and the problem was quickly fixed.

Another client, Hookle, aims to resolve client issues as quickly as possible. But sometimes, it’s hard for their support team to understand the problem a customer is facing. For instance, discrepancies occur between what a user describes and what actually happened. Smartlook’s session recordings make it easy to solve disparities and locate root problems faster than any other data source. 

Thanks to session recordings, their development team saves more than 10 hours a week locating and reproducing bugs — that boils down to a lot of hours saved each month.

In both actual accounts, you have enough user behavior analytics data at your disposal to detect and take quick action regarding any issues and problems your users may experience. 

Although there isn’t a need for comprehensive historical data in the above case, the ability to go back a month or 2 allows you to pinpoint the exact moment a problem first started. Such was the case with a quality assurance manager in an education-tech startup that used Smartlook to detect bugs under their radar for months.

It took their engineering team a long time to locate a bug affecting the desktop version of their app. They didn’t know that the metrics they were using to track the bug were misleading them.  

In actuality, the bug appeared on mobile devices when users accessed the desktop view in their browser. After using Smartlook to review session replays of the users who experienced the bug, the problem became apparent and they were finally able to correct it.

If you are determined to continually optimize the user experience of your website or app (which you should be), an extended data retention period will allow you to pinpoint precisely when an issue first appeared.

4. Identify trends in user behavior

If you aim to reduce churn, you need to rely on large amounts of data. This will help you spot patterns and identify trends that you can utilize to retain users. 

Retention tables are an excellent feature for spotting trends as they will help you understand how frequently and how long users engage with your website or mobile app after their first visit. 

Retention tables allow you to work with cohorts. In Smartlook, a cohort refers to a segment of users who share the exact date of appearance on your website or app. 

When analyzing a cohort, your data retention plan has a significant impact on identifying a user: a new user or a returning user. If a user returns on day 31, they will be seen as a new visitor if your plan stores data for 30 days. Depending on the type of business, user lifecycle, and the amount of traffic you receive, your data could potentially be skewed, leading you to form a wrong hypothesis and, subsequently, bad decisions. 

In Smartlook, you can define behavioral cohorts based on a specific action (aka event) a user performs on your app or website, like clicking on a sign-in or pay button within a particular time frame. Below is a cohort analysis of our trial users (users registered for a Smartlook trial plan). The ability to customize the date range of our analysis (for example, 3 days vs. 7 days vs. 10 days) allows us to receive precise user behavior data and decide the best course of action and – most importantly – when this should take place to convert them to paid users. 

 10-day retention table

Suppose you are a Saas company offering your users a monthly subscription. In that case, you can perform a similar analysis for, say, 30 days, 60 days, or 90 days to spot patterns relating to when users tend to become inactive or churn. From here, you can easily take the necessary action to retain them. 

Another way to gain insight is to compare one behavioral cohort, say free vs. trial users, to understand how long they remain active after performing particular actions. However, this is only possible when you have enough data to dive deep into user groups to observe the behavior that leads them to action (or inaction).

5. Spot patterns in user behavior

A great way to identify patterns is by breaking down Events based on properties. In Smartlook, breakdown tables provide a numeric breakdown and visual overview. They can also be paired with custom properties enabling you to tailor them to your specific needs. 

Custom properties allow you to concentrate your website or mobile app metrics into more meaningful data. A combination of properties and events will help you understand how often registered visitors buy from you, how blog visitors consume your content on specific days and times, and how certain user groups (e.g. pro plan users) interact with particular features.

Using my earlier example regarding analyzing the behavior of trial users, I can get a wealth of insight via custom properties, including “Is Setup Completed” and “Has Website Project” (see screenshots below). 

For our users to initiate a user behavior analysis with Smartlook, they will first need to add a piece of code to their website or native mobile app. This refers to the “Is Setup Completed” property. 

Additionally, when users log into Smartlook, they can select the type of project they wish to track — a website or a native mobile app. This information connects to the “Has Website Project” property.

With this information, let’s try to analyze the behavior of our trial users. Above, we see that 1,258 users installed the tracking code, with most users having a website. We also see that 1,120 mobile app users didn’t complete the setup. 

This high number tells us that there could be an issue preventing them from adding the snippet of code to their mobile app. With this information, I can form the following hypothesis: “Mobile app owners have difficulty implementing the code due to the nature of the asset.” 

Suppose I can confirm this hypothesis (by talking to customer support agents, the sales department, etc.) In that case, we could go ahead and create relevant content in our help center and even assign someone to onboard and guide them through the process.

The perfect data retention period

Unfortunately, there isn’t a “one size fits all” answer to this question. The important thing to remember is that the period for which you can store your data limits your ability to make data-driven decisions depending on your business goals. 

You can find an ideal Smartlook plan that suits your needs on our pricing page. You can choose a data storage period starting at 1 month (free plan), up to an unlimited number of months.

Sign up for a free Smartlook plan to better understand the data retention period that works best for you.

Renata Ekine
Renata Ekine

is the content marketing manager at Smartlook. She is a passionate digital marketer with experience in paid advertising, analytics, SEO, and lead generation. A data-oriented and creative-ideas seeker who loves creating engaging content.

0 %