An AppsFlyer attribution model, like the model in any mobile measurement tool, is a set of rules about which ad interaction gets credit for an install and how long after the interaction that credit can still be claimed. Those rules are tuned for apps where people install within hours. Service businesses sell over weeks or months. This article answers one question: how do you line up the windows and rules in your app tool with the attribution model on the levelpartner.ai platform, so the two reports describe the same buying cycle?
The three settings that decide the answer
Whatever tool you use, mobile attribution rests on three choices. First, which interactions count: a click on an ad, or also a view of an ad that was never clicked. Second, the lookback window: how many days back from the install the tool will search for a qualifying interaction. Third, the tie-breaker: when several ads were touched, which one wins, usually the last click.
Check your own tool's documentation for the options it offers and the defaults on your account. We do not describe a third-party product's settings here.
Why app defaults rarely fit a service sale
App tools are built for a short path from advert to install. A consultancy or estate agency buyer might see an advert, think about it for three weeks, ask a colleague, then install your app to book a valuation. A short lookback window credits that install to nobody, so it reports as organic. A view-through rule, meanwhile, can credit an advert the person scrolled past without noticing.
Meanwhile the platform records the first and latest touches it saw for each contact, with no window at all. The result is two models measuring different lengths of journey.
Before you start
- Your typical time from first enquiry to a booked meeting, taken from closed deals in the platform's pipeline.
- Access to the app tool's attribution settings, or the person who manages them.
- The platform's attribution fields populating correctly from forms and bookings.
- Custom contact fields for the app tool's attributed source and the interaction type, click or view.
Aligning the two models
- On your dashboard, click a Won Opportunities widget to open its records.
- Use Export and work out the median days to a first booked meeting.
- Compare that figure with your app tool's click lookback window.
- Agree with the app tool's owner whether to lengthen that window.
- Decide in writing whether view-through credit counts in revenue reports.
- Ask the app tool's owner to send click or view with each event.
- Store that value in a custom contact field through your webhook workflow.
- Edit an Opportunities widget, open Conditions and click Add Filter.
- Select Attribution and pick First or Latest Attribution to match the app tool.
- Click Save, then Save Changes.
The point of view worth holding
View-through credit is the setting most worth arguing about. It makes awareness spend look productive, which it sometimes is, and it inflates the numbers of whichever network serves the most impressions. For budget decisions in a firm with flat revenue, count clicks, and report views separately as a supporting signal. You give up some credit for genuine awareness work. You gain a report that cannot be inflated by impression volume. The multi channel attribution guide covers how to weigh awareness channels instead.
Say the median time from first touch to a booked valuation is 19 days, and the app tool's click window is 7 days. Over a quarter, 40 app bookings might report as organic. Check them against the platform's records and suppose 22 first arrived through paid social more than a week before installing. A window closer to 19 days would credit most of those to paid social instead. The figures are illustrative.
What changes for the business
Once both models use similar windows and the same tie-breaker, the app report and the CRM report agree closely enough to plan with. Marketing stops defending numbers the sales team does not recognise, and budget moves on evidence both sides accept.
Time to maintain
The alignment work takes a morning. Repeat the cycle-length calculation every six months, because the buying cycle changes as your offer and market change. Weekly upkeep is nil once the fields are flowing.
Teams copy the app tool's default windows into every report and then use them to judge channels with long consideration. Paid social and content suffer most. Always set windows from your own closed deals, not from defaults.
When not to change the app model
If the app is mainly used by existing clients, the acquisition model hardly matters, so leave the defaults alone. And if app installs mostly happen immediately after a click on a booking advert, the default short window may already fit. Measure before changing anything. For the general choice of model, read choosing an attribution model.
Troubleshooting
Organic installs look suspiciously high. The lookback window is shorter than your buying cycle. Compare app-organic contacts against their earliest platform touch to confirm.
One network claims most installs. View-through credit is probably on and that network serves the most impressions. Report click and view credit separately.
The platform and the app tool still disagree. One is using first touch and the other last touch. Match the tie-breaker before comparing.
Frequently asked questions
What lookback window should a service business use?
Set it from your own data: the median days from first touch to conversion on closed deals, with some margin. Many service businesses find their real cycle is several times longer than app defaults.
Is view-through attribution worth using?
As a secondary signal, yes. As the basis for budget, it tends to overstate the networks that buy the most impressions. Keep it in a separate column.
Does the platform use lookback windows?
The platform stores the first and latest recorded touch for each contact without expiry, and contact and opportunity widgets can filter on either. That makes it a useful neutral reference when you check the app tool's model. For the full dashboard layout, see the marketing attribution dashboard guide.