Starting product discovery

J. Kelly Brito

1. Starting discovery ←
2. Feedback database

When you start a new product management role, you don’t have any first-hand evidence for customer needs yet. Talking to customers remains the gold standard for product discovery, but it takes time for it to expose a full range of customer problems, and other product opportunities.

Use the last few months’ data from existing sources, both to bootstrap a product feedback database, and to surface questions to follow up with customer research.

Customer support requests

To start with, make sure you know about any critical urgent issues. Your customers will tell you about anything that blocks them, possibly daily, so start there. Get access to incoming customer support requests, so you can read customers’ frustrations in their own words.

The most obvious issues occur when a feature doesn’t work (bugs), or when it works but customers require help to use it (UX issues). Occasionally, the product can’t do what the customer wants, in which case you need to figure out what they wanted to do.

Bonus points: automatically forward the (anonymised) contents of each new support request to a Slack channel, to get an early signal about customer-facing issues, and customer needs.

Sales & marketing

In a B2B context, your sales team can tell you what would make the product easier to sell. They probably already have a spreadsheet. In a B2C context, the marketing team’s social media manager can share the consumer equivalent, based on social media posts.

These sources resemble support requests, but with a different flavour. In both cases, they may expose broad patterns, but lack specific evidence. Instead, they indicate potential topics to explore further. Use these sources to collect hypotheses that require evidence, direct from customers, and then follow-up with those customers, to discover the full story.

Bonus points: ask sales/marketing to segment their list of topics by persona, so you get evidence of who cares about each topic.

Product analytics

In one of my product manager roles, a discussion with a UX designer led us to review the product’s behavioural analytics. We thought that we might find some clues about how we might improve user experience. As it turned out, we only had to watch ten session replays to witness an obvious issue that affected many people, and which we immediately wanted to fix.

Sometimes, you can learn a lot from behavioural analytics, if you have the right tools. Depending on how the product works, and the data you collect, you might discover:

• user frustration, such as dead clicks (that do nothing), and rapid navigation
• task failure – not completing multi-step tasks
• user errors – not using the product the way you designed it to work
• system errors – trying to use the product when it doesn’t work.

Bonus points: publish an automatic activity feed for serious issues, to create visibility of the scale of problems, and motivation to fix them.

Other channels

An existing product will always have existing recent feedback from customer support, sales, and marketing. With luck, as well as product analytics, you’ll have access to product feedback from other existing channels:

• in-app product feedback
• in-app surveys
• churn surveys

Bonus points: publish a combined feed from all of the above channels, so all product team members can stay in touch with what customers say.