👉If you liked this, you should sign up for my newsletter! I share helpful resources like this each week.

Find overlap between features to build smarter user journeys

📎  Sample data set:

Expanded_Collaboration_Platform_Dataset (2).csv

Analyze the data

I have a dataset containing user engagement counts for various attributes across multiple features of a product. Each row represents a unique combination of attributes, such as:

Attribute Type: Categories like Role, Team Size, Day of the Week, Region, Platform, Session Frequency, etc.
Attribute Value: Specific values for each attribute type (e.g., Manager, Small Team, Monday, North America, Web, Daily).
Columns for feature engagement counts, such as Feature A, Feature B, Feature C, etc.
Objective:
I want to identify which features (e.g., Feature B, Feature C) share the most similar user base with a target feature (e.g., Feature A), based on the distribution of attributes.

Steps to Analyze:
Calculate Similarity Scores:

Use cosine similarity or Pearson correlation to calculate how similar each feature is to the target feature across all attributes.
Attribute-Level Insights:

For each attribute type (e.g., Role, Day of the Week, etc.), calculate similarity scores between the target feature and each other feature.
Highlight which attributes (e.g., Platform or Region) show the strongest alignment between the target feature and each other feature.
Rank Features:

Rank the features based on their overall similarity to the target feature.
Highlight any patterns in attributes that contribute to high similarity scores.
Provide Recommendations:

Based on the analysis, identify which features should be prioritized for promoting the target feature.
Suggest specific strategies for building user journeys (e.g., targeting users in specific roles, industries, or platforms).
How can I use this analysis to design targeted user journeys to guide users from high-similarity features into the target feature?

Output Format:
A table showing similarity scores between the target feature and other features for each attribute type.
A ranked list of features by overall similarity to the target feature.
Recommendations for targeting and user journey design.

Build user journeys

Give me 3 extremely specific recommendations for user journeys I could build from the most similar feature that accounts for the specific ways in which it is similar. Include specific copy recommendations targeting the recommended audience. Explain your rationale

For reference:
1. AI meeting assistant automatically summarizes meetings, generates action items, and shares follow-up tasks with the team
2. Task manager helps team organize and assign tasks, track progress, and hit deadlines.
3. Document collaboration allows teams to edit, comment, and collaborate on documents in real time.
4. Team Pulse is a feature that helps managers collect feedback and track team sentiment using quick surveys.

Define the right activation metric (and move it!) with App Store reviews

Generate data set

Obviously if you have access to real app store reviews, you should use those. I created a synthetic data set for my demo using the below prompt. If you just want to try out this technique you can do the same.

give me a sample data set of 1000 app store reviews for the headspace app. they should reflect real reviews in both their praise and their criticism, though with more specificity about specific features, experiences, or UX patterns than a typical app store review might include

Select activation metric