Validate your ideas fast and cheap

In this episode of Datadrivet, Joni Lindgren and Jasmin Yaya talk to Tom Airaksinen, Product Owner at PE Accounting, ahead of his Product Management Festival talk “Scrappy Mixed Methods: The Resource-Strapped PM’s Guide to Getting Data-Informed.”

Mixed methods means combining quantitative and qualitative ways of gathering insight. Scrappy means getting real work done without expensive tools or senior specialists. Tom came up as a UX designer focused on qualitative work, then picked up quantitative methods and moved toward product management. One recommendation he makes for anyone in product: learn SQL so you can query the database yourself.

He frames the Product Owner role with a Crisp definition: the main job is “maximizing the amount of work not done.” When time and resources are tight, deciding what not to build is the work.

The PE Accounting app gives the concrete story. Tom saw that user needs around expense and travel logging were not met by the existing features, so he built a simple prototype to test the hypothesis. During the pandemic the team pivoted toward manager approval (expense attestation) after user research pointed there. A 25-minute Teams interview dug into why managers wanted this on mobile, and the answer was bigger than expected: push notifications let a manager approve an expense without checking work email on vacation.

To get scale on the signal, Tom ran Hotjar popup surveys on relevant pages and collected hundreds of responses in days. One lesson came from a question that backfired: “Would you like to approve/attest invoices via mobile?” was double-barreled, asking two things at once, so the answers were hard to trust.

For sorting feedback, Tom uses three buckets: dismiss it as an anomaly, flag it for more investigation, or accept it as valuable.

The takeaway is that you do not need budget to be data-informed. Cheap prototypes, short interviews, and quick surveys get you most of the way, as long as you ask clean questions and stay honest about what each signal means.

Listen to the full episode of Datadrivet for Tom’s full account of validating ideas without a big toolset.