Right after its launch, we started getting the anticipated signals from users and interpreting them. So the team was able to
readjust the product and marketing strategy several times before getting to the product-market fit phase.
Once we got our
initial traction, the team threw all its resources toward user activation. Having both behavioral and qualitative data in hand
through a series of meticulously designed experiments, we reached a 2.5x increase in user activation rates.
First of all, we established the Net Promoter Score collection and built multiple flows to work with the user feedback. One of the
simplest but super effective exercises was streaming individual NPS responses, with an anonymous view of user actions within the
product, to a separate Slack channel. Over the course of several weeks, nearly every team representative – product, operations,
engineering, and support – was discussing each of the issues, one by one. Also, with the NPS, we made our happy users finally speak
up. Their sentiment provided powerful input for the new marketing campaigns.
At the same time, we sought to identify
loyal audiences by leveraging mParticle. Once we identified multiple personas of power users, we helped the marketing team increase
the share of those user groups via audience/lookalike building, which we saw, in no time, being reflected on the Average User
Lifetime Value and Return on Advertising Spend as a result.
To address compliance demands, we built a secure automated solution to handle all policies related to in-storage data. At the same
time, this implementation has allowed the support team to focus on the customer and leave any privacy-related requests to automated
During our three years working with the client, we've hit some rough patches along the way. We've
really maneuvered around all kinds of potential early-stage startup mistakes. The primary challenge has been to reduce the high cost
of user acquisition. New privacy regulations have made this effort much more problematic.
We've also encountered
multiple issues with a set of recently developed manual pipelines. However, this has only fueled our motivation to change the game.
We switched to Fivetran, bringing data quality to a new level, and oh, as a bonus, our engineers were able to sleep better.