--- title: "What product data cannot tell you: Maria Petrova, ValueLab | Skene" description: "ValueLab's Maria Petrova on why a metric answers what but never why, why a service blueprint is not the customer journey, and why tracking starts on day one." url: https://www.skene.ai/resources/the-question/maria-petrova doc_version: "c00117dbc7a2" last_updated: 2026-10-30 --- Resources / The Question / Issue 5 # Maria Petrova: data tells you what happened, never why Maria Petrova. Co-founder of ValueLab, a Helsinki product management consultancy. Issue 5 of The Question, published by A Founder's Friend on 12 August 2026. By [Nicolas Dolenc](https://afoundersfriend.substack.com), author of A Founder's Friend. Skene adapted this page from his published issue. Founders sometimes tell Maria Petrova they have 50 customers and no need to track anything yet. After more than fifteen years leading product at Smartly.io, Zalando, Supermetrics and TWAICE, she disagrees, and her reasons have little to do with dashboards. **The short answer:** Data establishes what happened and never why, and tracking is infrastructure to build before the first real user, not something to add later. ## What Maria Petrova trusts, and for what | Evidence | Trusted for | Not trusted for | | --- | --- | --- | | Product metrics | What happened | Why it happened | | A direct call, in B2B with few accounts | Resolving a hypothesis fast and cheaply | Consumer products, where one call is one unrepresentative opinion | | Structured tests, with thousands of users | Resolving hypotheses at scale | Small B2B bases, where a direct call settles a hypothesis faster and more cheaply | | The company's map of screens and clicks | How the company wants users to move | The customer's real journey, which starts before the product opens | ## The job is de-risking ValueLab places senior product leaders inside client companies to own the work and ship it. Maria frames product decisions around one constraint: a user who tries a bad solution once rarely gives a second chance. The goal is not to follow a framework. It is to raise the odds that the first real attempt meets a real need. ## What, never why A metric showing that a feature goes unused does not say why. It sets up competing hypotheses, a visibility problem or a motivation problem, and a person has to resolve them by asking, testing or watching. Which method is right depends on the customer. With a handful of enterprise accounts, call them. With thousands of consumers, one call is one opinion, so run a test. > data tells you what happened, never why, and the “why” still needs a human. A Founder's Friend, summarising Maria Petrova, in Issue 5 ## Your map of the product is not the customer's journey Many teams mistake a service blueprint, the screens and clicks they designed, for the customer journey. The blueprint is how the company wants things to go. The real journey starts earlier, with the problem and the moment someone realises they need help. Her example is a marketing lead who wakes up sensing something is wrong in the campaign data and has to click through five dashboards to find out what. The opportunity is not a better version of those five. It is one place that answers fast, because how quickly the answer arrives is what makes that customer happy. ## Why tracking starts before the first user Her closing point was unprompted. What users say they will do and what they do diverge. Building with AI assistance makes shipping fast without making it self-documenting. And tracking only gets harder to add the longer a team waits. > Tracking is not optional context to gather someday; it’s foundational infrastructure to build in from day one, before the app even has its first real user. A Founder's Friend, summarising Maria Petrova, in Issue 5 ## What to do with this When a metric moves, write down the two or three hypotheses that could explain it before anyone picks one. Then choose the cheapest way to separate them: a call if you have few customers, a test if you have many. For seven checks drawn from across the series, see the [questions to ask before trusting a number](https://www.skene.ai/resources/the-question#before-you-trust-a-number). ## Frameworks this interview tests - [Jobs to be done (JTBD)](https://www.skene.ai/resources/frameworks/jobs-to-be-done): Clayton Christensen's jobs to be done says customers hire a product to make progress. The job decides which events matter, so interviews come before the tracking plan. - [Aha moment and time to value](https://www.skene.ai/resources/frameworks/aha-moment): The aha moment is the action after which users become far more likely to stay. A borrowed number misleads; your own needs a validated event and a fixed time window. - [North Star Metric](https://www.skene.ai/resources/frameworks/north-star-metric): A North Star Metric is the one number that captures the value customers get. It is usually built from several events, so a break in any one of them moves every team. ## Read next in The Question - [AmirBahador Bahadori: product data is three layers, not one](https://www.skene.ai/resources/the-question/amirbahador-bahadori) - [Dino Repo: one metric, five numbers, and a ten-minute fix](https://www.skene.ai/resources/the-question/dino-repo) - [Niko van Eeghen: one metric, three meanings at Evondos](https://www.skene.ai/resources/the-question/niko-van-eeghen) - [What all fourteen interviews add up to](https://www.skene.ai/resources/the-question) ## Source and disclosure [Read Issue 5 on A Founder's Friend](https://afoundersfriend.substack.com/p/aff-the-question-issue-5-making-decisions). The full interview is for subscribers. This page paraphrases it and quotes nothing that sits behind the paywall. Skene sponsored this research and commissioned it on a single question. A Founder's Friend ran the interviews, wrote them up and published them, and the series states that Skene paid for the research and did not direct the questions, the analysis or the conclusions. Each of the thirteen interviews was reviewed and signed off by the person in it, named or anonymous by their choice. Appearing in the series is not an endorsement of Skene, and nobody here is presented as a Skene customer. ## Sitemap - [Markdown sitemap](https://www.skene.ai/sitemap.md): every page on this site, with its markdown counterpart. - [Index for language models](https://www.skene.ai/llms.txt): the same set as a link-dense index. - [Full corpus](https://www.skene.ai/llms-full.txt): reference content concatenated as plain text.