Missed add-ons, location by location
Sample dataHow to read this
- Customers often book two services together.
- The booking site can suggest the second at checkout.
- When it never does, that is a missed booking.
- Suggestions customers rarely take can be swapped out.
- Each dot is a service.
- A line = booked in the same visit. Thicker = more often.
- Dashed line = never suggested. That is the opportunity.
- Select a dot to see its suggestion.
- Hair
- Colour
- Nails
- Skin
- Massage
- Brows and lashes
- Thicker: more often
- Dashed: never suggested
- Ringed: nothing suggested with it
Sample data. The real project is under NDA.
A close replica of the map with sample data - the real project is under NDA, so the data and figures here are an example.
The problem
The business knew a lot about what its customers do, but the answers sat in separate systems, so simple questions could only be answered one spreadsheet at a time.
The approach
Join the data into one model and treat it as a network: what gets booked together becomes a line between two points, so the add-ons customers already choose - but the booking site does not offer - stand out, along with the few that rise with the seasons.
What was built
- One clean dataset. A repeatable pipeline that pulls the sources together.
- Suggestions, not decisions. Add-ons to suggest, ones to drop and seasonal notes, each waiting for a person to approve.
- A map to explore it. The interactive view above, so the team can see the patterns rather than read them in rows.
Seasonal add-ons
Most add-ons hold steady all year. A few rise with the seasons - those are worth featuring ahead of their busy months.
- Pedicure with Gel manicureBusy Jun to AugFeature from May
- Updo with Gel manicureBusy Nov to DecFeature from Oct
- Hot stone massage with Signature facialBusy Nov to FebFeature from Oct
Illustrative sample data, not real figures.
Results
One clean model replaced the hunt through spreadsheets. The team can now see what customers book together and which add-ons rise with the seasons, and acts on suggestions grounded in real booking patterns. Since working together, the average cart value of the business has risen by approximately 10%.
Inside the analysis
- Pairs counted across every past booking
- A check that each pair is more than coincidence
- Pairs seen too rarely left out
- Month by month checks for seasonal pairs
- A person signs off every suggestion
What the team gets
- Add-ons to suggest at checkout
- Suggestions to drop
- Seasonal add-ons to feature
- An interactive map to explore
Could this be you?
- If your data already knows what customers buy together but nobody can see it, this is the pattern that turns it into decisions you can act on.
Next case study
Four times the clicks from Google