Upaj farmer experience

Day-14 and Day-30 retention had stayed flat for thirteen months. Monitoring a farm, checking mandi prices and finding relevant information were recurring tasks, but usefulness also depended on reaching them from a constrained phone.

We organised the revamp around observing the farm, checking mandi prices and finding relevant information. These are reasons to revisit a product; a visual refresh or an increased number of prompts would not by itself establish that usefulness.

The work brought monitoring, market information and a new notification structure into the experience. The records document that combined scope, rather than a tested comparison between these alternatives.

Useful information still had to be accessible on constrained Android phones. Later portfolio material describes entry reducing from eleven steps to three, with attention to app size and cold-start behaviour. These are concrete access changes, not a separately measured onboarding-conversion result.

That same later material describes crop-cycle notification triggers. Notifications were a route back to a relevant task, alongside the content and experience work; they were not the sole product strategy.

The career record reports improvements over the programme’s first three months. Several experience, content and communication workstreams changed together, so these are programme reports rather than controlled estimates of any one feature’s effect.

For another retention problem I would begin with the task worth repeating, then inspect the effort between the user and that task. In Upaj, farm information, market prices and content gave that investigation a concrete starting point; lifecycle messages helped people reach it.

Rural creators and influencers contributed to the content library. Material longer than a minute used YouTube; shorter formats used a custom CDN. The decision connected content length and distribution to the mobile experience.

I coordinated the content and information work with design, engineering and lifecycle communication. The record does not provide a feature experiment showing that one content format or notification caused the retention change.

The eleven-to-three-step entry change and app-size/cold-start work are recorded in later material. Historical before/after screens and the exact eight removed or deferred steps are not supplied. The case therefore explains the documented access problem without fabricating a screen sequence.

Monetisation remained in my remit, but no separate revenue outcome for the entry, content or notification changes is established.

Farm OS monitoring and crop-health diagnosis are separate product studies. This historical capability map records their relationship to the broader farmer experience; it is not evidence of one shared deployed backend or a measured engagement effect from every component.

The table is retained as background. The engagement account concerns entry, useful tasks, content and return; crop datasets, model quality and sensing architecture belong with the related technical cases.

I would first reconstruct activation and repeat-use cohorts, then check whether improvements held across target phones and network conditions. Those are proposed follow-up questions, not historical experiment results.

Other self-reported programme measures in the career record are active users from 100,000 to 600,000, CAC falling 66% and average sessions from eleven to fifteen minutes. Active-user definitions, comparison periods and raw cohorts are unavailable. These figures are retained as source context, not promoted as headline proof.