The castle of Zendesk and the glowing crack

Platforms are castles, and the Zendesk App Marketplace was no exception. We often look for subtle indications of how these digital fortresses direct attention - the unseen mechanisms that govern what users see, and what remains hidden. In this castle, we found a glowing crack in the walls of its internal search engine.

We noticed that the marketplace's search algorithm seemed to favor newness. Apps that had been recently updated, even with minor changes, often appeared higher in relevant search results than older, even more established, alternatives. This wasn't explicitly stated in any documentation; it was a pattern we started to piece together through observation. Attention was leaking out of the seams of static, unchanging app listings, flowing instead towards those that showed signs of recent activity and improvement.

What we noticed about the flow of attention

The Zendesk App Marketplace serves millions of users looking to extend their customer service capabilities. For app developers, visibility within this ecosystem translates directly into adoption. We started by observing how various apps, including our own, ranked for core keywords. We logged their update dates, descriptions, and current positions. A consistent trend emerged: apps that had received recent updates often held more prominent positions, especially within the first few pages of search results, which capture the vast majority of user clicks.

This wasn't about major feature releases; we saw smaller, maintenance-style updates also contributing to this recency bias. It became clear that the platform's internal logic was, perhaps inadvertently, signaling to users that active development equated to a more reliable or relevant solution. This was our crack - a way to signal vitality directly to the algorithm.

The mechanic: regular, targeted app updates

We decided to run a specific Growth Hack over a 60-day period. Our goal was to leverage this observed recency bias to increase our app's visibility and drive installs. We formulated a plan that involved a consistent schedule of minor updates and careful keyword optimization.

First, we identified the key search terms our target users were employing. These were specific problem-solution queries that our app directly addressed. Next, we began a cycle of app updates, not necessarily tied to new feature development, but rather to maintenance and optimization. Each update, pushed every 7-10 days, included minor code refinements, dependency updates, or even slight adjustments to the app's metadata - its title, description, and keywords. We ensured these changes were legitimate and within the marketplace's rules, never compromising the app's stability.

For each update, we refined the app's title and description to subtly incorporate a broader set of relevant keywords, ensuring they read naturally to users but were also algorithmically optimized. We tracked our app's search ranking for these keywords daily, alongside the number of new installs.

What happened: a consistent upward trajectory

The results were concrete and immediate. Over 60 days, our app garnered more than 200 new installs. We saw a clear correlation between our update schedule and our search ranking. Within 24-48 hours of pushing an update, our app’s visibility for target keywords consistently improved, often jumping several positions to appear on the first page of results. This uplift in visibility translated directly into an increased click-through rate to our app’s listing page, and subsequently, a higher volume of installs.

The install rate peaked after each update, then gradually tapered off until the next update refreshed its ranking. This created a 'sawtooth' pattern of install acquisition, with a clear upward trend over the two-month period. We also noticed that competitive apps, especially those that remained static for weeks or months, slowly slipped in rankings as we maintained our consistent update cadence. This confirmed that attention was indeed leaking from their stagnant listings and flowing towards our more 'active' presence.

Where the crack held and where it needed reinforcement

The core mechanic - consistent, minor updates - held remarkably well. It proved to be a legitimate and effective way to influence search visibility on the Zendesk App Marketplace. We maintained an average of 3-4 updates per month, and this frequency appeared to be optimal for sustaining an elevated ranking without over-saturating the platform with 'dummy' updates, which we strictly avoided.

Where we noticed limitations was in the depth of keyword optimization. While titles and descriptions had an impact, the marketplace's algorithm likely considered broader factors beyond mere text matching. Further refinement of our app's category and tag selections might have amplified the effect. Additionally, while the raw number of installs was strong, understanding the long-term retention of these users became the next frontier. We gained the installs, but ensuring their sustained engagement is a separate challenge.

What we are running next

Next, we are exploring the impact of localized app descriptions and titles across different Zendesk marketplace regions, seeking to capture attention from non-English speaking users who may use different search terms.