A contextual upsell is an in-app offer triggered by what a user is doing at that moment, rather than a fixed message everyone sees on the home screen. It matters because the timing carries part of the argument before the copy does.

The EleFant, an Indian toy-rental subscription app, ran one on its toy-view screen. A parent already comparing a specific toy got the membership pitch while they were weighing the decision. According to Digia's September 2026 EleFant case study, that single placement hit a 36.6% click-through rate, the highest in a portfolio of 10 live experiences.

The Max-plan banner on the home screen ran at the same time, reached far more people, and did not come close. The upsell met interest the user had already shown through their own behaviour, and gave them a reason to widen a decision they were in the middle of making.

Across all 10 experiences: 60,869 impressions and 11,337 clicks, a blended click-through rate of roughly 18.6%.

That blended figure tells you the portfolio worked as a set. The gap between 36.6% and 18.6% tells you where user intent was already sitting before any campaign showed up. That is where the next round of experiments belongs.

Advance booking already worked inside The EleFant app before any of this started. Nothing pointed users to it at the moments it was useful, so it sat there.

The fix was surfacing it on the home screen, the order-history screen, and the all-toys screen. Those placements produced 299 completed advance bookings, a downstream action rather than a click count. No feature work, no release.

Inside the same 30-day term, The EleFant's growth team built and launched its own app-education video campaign without vendor involvement.

Easy to skip past, because there is no metric on it. It also decides whether the numbers above repeat next month. A team that files a ticket for every idea runs four experiments a year rather than ten in a month.

The full case study breaks down all four placement strategies The EleFant ran in parallel, the ramp from 5 to 10 live experiences, and six questions on why the narrow placement beat the broad one.

Push a saved campaign from the dashboard to a connected debug device before anyone sees it. Scan the QR from Test & Debug, enable Sync on the device, then send any draft or unsaved edit and watch it render on the actual screen it will fire on.

The dashboard reports status per device: Sent, Received, Shown, Failed, or Unconfirmed, with a failure reason when the campaign does not render. Missing tooltip anchors and unfilled variables are the two usual culprits.

One constraint worth knowing upfront. Live testing runs on debug builds only. Release builds will not connect as test devices.

Read the live testing guide → Live Testing