Callback Tracker
Finds the return trips hiding in your book - a completed job, then another job at the same property inside your window - lays the pairs out for your verdict, tags the confirmed callbacks so they stay found, and reports your callback rate by technician and by month from your own tagged history. No benchmarks, no accusations: you judge every pair.
The problem
Every shop eats return trips, and almost none can say how many. There is no callback field in Housecall Pro - the truck rolls twice, the second visit bills zero or small, and the book swallows it as just another job. The rate nobody wants to know stays unknown.
And the data alone cannot answer it: a callback and new work sold on a return visit look identical in the records. Somebody who knows the jobs has to judge - so this skill makes the judging fast instead of pretending it is not needed.
What it gives you
- Candidate pairs, strongest callback smell first: a completed job, then another job at the same property inside your window - the return that billed zero or near-zero, the days between, same technician or different, both descriptions side by side.
- Recurring maintenance visits and canceled jobs excluded by rule; same-day pairs held separately as possible split billing rather than mixed in.
- Your verdict on every pair - callback, not a callback, cannot tell. Nothing unconfirmed is ever counted in the rate, and nothing is ever tagged without you.
- Confirmed callbacks tagged on the return-visit job so the answer accumulates - next month reads its own history from the tags.
- The rate report: confirmed callbacks over completed jobs, by technician (attributed to the original job's tech), by job type, by month - with denominators shown and thin ones called thin.
What you need
| Tier | API Key Required |
|---|---|
| Reads | The MCP |
| Needs | The free MCP plus a Housecall Pro MAX-plan API key you generate. Applying changes needs a full-access key. |
| Re-run | Monthly. Each run's tags make the next one's history. |
This one makes changes
Adds a tag to jobs you confirmed, one at a time; removing it is one ask away.
Every change is confirmed individually with the before and after shown. There are no batch approvals, because on a live book that is how a small mistake becomes a large one.
Your first ask
Once it is installed, this is the literal thing you type:
What it cannot do
- It cannot tell a callback from new work sold on a return visit - nobody can from the data, which is exactly why you judge every pair.
- The only write it performs is the tag you confirmed; removing one is one ask away. Tags are internal - nothing about tagging reaches any customer.
- No industry benchmark is ever quoted. Your own trend is the honest comparison, and it takes two or three runs to exist - the first run says so.
- Requires the MCP (v1.15 or newer) and a MAX-plan API key. A read-only key covers everything except the tagging step.
Updates
Every purchase includes all future updates to this skill. Your book decays continuously, so the same job is worth doing again, and each version gets better at it.
The skill checks this site for a newer version when it runs. That request carries the product name and the installed version, and nothing else. There are no account details, no data from your book, no telemetry of any kind. If the check fails for any reason it says nothing and carries on.
Buying it
Sold in the Team & Quality pack: $39 one time at half price while the shelf is new for all 2 skills in it. The pack is the unit because a home service company that needs this one usually needs the rest of it the same week. What is in the pack. Lifetime is $149 and covers every pack on the shelf plus every skill added to any of them afterwards.
Questions about whether this fits a particular book go to hello@hcpapi.com , including the question of whether it is worth buying at all.
Independent project. Not affiliated with, endorsed by, or sponsored by Housecall Pro.