For most companies, power users are the driving force behind the success of many digital products, wielding significant influence and engagement compared to the average user. They are not only highly active and skilled with the features of a product, but they also often serve as brand ambassadors, providing valuable feedback and promoting the product within their networks.
Identifying these key players is crucial for any business looking to leverage their insights for product development and growth. By experimenting on power users, companies can fine-tune their offerings better to meet the needs of their most engaged customers, ultimately driving retention and attracting new users. I'll be covering the characteristics that define power users, explore strategies for pinpointing them within your user base, and discuss the benefits and methodologies of conducting experiments specifically tailored to this influential group.
What are some characteristics of Power Users? The true definition depends on your company/business model. However, I believe that there are some common themes across verticals.
1. Power users are distinguished by their deep engagement with your product. They are the ones who use advanced features, spend more time within your application, and/or complete more important actions than other users.
For example, we may measure a power user at Statsig as having created more than 25 experiments within the last 3 months. The data helps you determine the power users. Quantitative.
2. They are constantly providing constructive feedback to either a product feedback form or an account team if your business is B2B.
For example, at Statsig we have a customer advisory board and are in constant communication with customers via interviews and Slack channels. Qualitative.
But what set of actions actually determines a power user? We reviewed some of the indicators for a power user for each product line at Statsig, which include:
15 Feature Gates Created in the last month
Average of 10 queries run through our analytics product per session
25 Experiment Created in the last quarter
25 Session Replay Videos Viewed in the last month
Depending on your vertical, these will differ, but the common denominator is the usage or stickiness of the product in some sense.
In E-commerce, usage may include # of purchases made weekly, in the travel industry, this may be the # of trips booked yearly. Usage or stickiness is key. In order to understand the threshold that may determine a power user, you need to have a good baseline measurement of your users.
Our head of product, Margaret-Ann Seger, wrote an in-depth blog on the suite of out-of-the-box user accounting metrics that Statsig provides to make it easy for growth teams to measure and optimize user behavior that can help you get started building that baseline if you haven’t already.
Now that we have decided who our power users are, let’s give an example of how to build out an A/b test against these users/companies.
Before experimenting, ensure you create a Statsig account and have integrated your SDK into your product. Familiarize yourself with the platform's capabilities and set clear objectives for what you aim to achieve through experimentation.
For a list of all SDKs, check out the docs here. I will just use some boilerplate pseudocode to give examples of passing data through the SDK and target that group within a Statsig experiment.
When segmenting users within Statsig, it is a good idea to understand which product you are using; these setup steps may look a little different. We offer two products for experimentation. Cloud and Warehouse Native.
When using our Cloud product, targeting user metadata is done through the user object’s custom properties. To set these up, check out our docs here.
When using our WHN product, you can also cohort against Entity properties found here.
Statsig allows you to segment users based on attributes like behavior, demographics, or custom properties. Consider attributes that reflect their high engagement levels similar to the examples shared above. This might include frequency of use, feature adoption, or even customer lifetime value.
When segmenting users, aim for:
Clarity
Relevance
Consistency in user attributes across platforms is crucial for accurate segmentation. This granularity enables you to target specific user groups, like power users, for more focused experiments.
Here are some code examples of how to accomplish this. For the examples below, we will cover the process for Statsig Cloud.
In this example below, we will use an example similar to how we might define a power user of Statsig. This is using a property checking how many experiments the user has created in the last 6 months.
lastSixMonths = (customerId) => { count = 0 // write some database query to fetch number of experiments created in the last six months. return count } ... statsigUser = { userID: 'a-user', custom: { // This data is completely optional, and can be any data available to your application customerType: "premium", experimentsCreatedLast6Months: lastSixMonths("a-user"), } }
After we have coded in how to represent a power user within the User Object, lets now take this data and pass it into a rule within an experiment. Here in this screenshot, we are creating a rule that is targetting a custom property we called <experimentsCreatedLast6Months>. We are saying that if this company had created 25 or more tests in the last six months, it would have been considered a power user of Statsig.
Now, let's say that we have our definition, and we want to share this rule or config across multiple experiments or flags in a scalable way. We can do this through a Segment in Statsig.
We can then reuse this segment in any test or feature gate we choose. This means that setting up a targeted test doesn’t require engineering to verify the condition properly, reducing the risk of shipping a broken test or feature.
If the ultimate goal for your business is to build a product that solves a problem your customers have, who better to help you shape that vision than the ones most invested?
Power users are an extremely viable option when it comes to product feedback requests, early beta testing etc. I talked about the characteristics that define power users, explored strategies for pinpointing them within your user base, and discussed the benefits and methodologies of conducting experiments tailored explicitly to this influential group. Understanding and segmenting these users allows you to take action quickly and build products your entire user base will love.
Kayak reacted quickly to news coverage of airline-related catastrophes and gave its Aircraft Filter feature more visibility, resulting in a 15x increase in user engagement.
A/B testing is the most reliable way to get evidence. Whether you're an advanced experimenter, or delving into testing for the first time, here's what you should know:
The new Pluto design system is an effort to simplify and enhance your experience with Statsig. Take a look for yourself!
Explore technical strategies for scalable experimentation, focusing on overcoming information and managerial challenges to achieve sustainable success.
Why do analytics teams fail? It often pertains to communication, collaboration, and aligning data efforts with business goals. Watch the video:
As our product matures and our customers use Statsig to do more, the goal of the default landing experience evolves. Find out how the Home tab has evolved to support this.