Product Analytics: What Marketers Keep Getting Wrong

Product analytics is the practice of measuring how users interact with a product, tracking behaviour across features, flows, and sessions to understand what drives engagement, retention, and revenue. Done well, it closes the gap between what marketing promises and what the product actually delivers. Done poorly, it produces dashboards full of activity data that nobody acts on.

Most marketing teams treat product analytics as someone else’s job. That’s a mistake. The data sitting inside your product is often the most honest signal you have about whether your acquisition strategy is working, which segments are actually valuable, and where the customer experience breaks down after the click.

What Product Analytics Actually Measures

There’s a version of product analytics that gets confused with web analytics, and it’s worth separating them clearly. Web analytics, the kind you get from GA4 or similar tools, tells you about traffic: where it came from, which pages people visited, how long they stayed. Product analytics goes deeper. It tracks specific user actions inside a product, feature adoption, flow completion, drop-off points, and how behaviour changes over time for individual users or cohorts.

The distinction matters because the questions are different. Web analytics asks: did people come, and did they convert? Product analytics asks: after they converted, what did they actually do? Did they use the feature you built? Did they come back? Did they reach the moment where the product becomes indispensable to them?

In practice, the tools overlap. GA4 has event tracking that can approximate some product analytics. Dedicated tools like Mixpanel, Amplitude, and Heap go further, with user-level tracking, funnel analysis, and cohort retention built in as core features. The right choice depends on the complexity of your product and what questions you’re trying to answer, not on which tool has the most impressive demo.

If you’re building out your analytics capability more broadly, the Marketing Analytics hub at The Marketing Juice covers the wider measurement landscape, from attribution to GA4 implementation, with the same commercially grounded approach.

Why Marketing Teams Need to Care About In-Product Data

I spent a chunk of my career running agency teams focused on paid search and performance channels. We were good at driving traffic. We could optimise cost per click, improve quality scores, and build campaigns that generated volume efficiently. What we were less good at, and this was an industry-wide blind spot, was understanding what happened after the click.

At lastminute.com, I ran a paid search campaign for a music festival that generated six figures of revenue within roughly a day. The economics looked brilliant on the channel dashboard. But the honest question I should have been asking was: how many of those buyers came back? How many created an account? How many became repeat customers? The acquisition metric was clean. The downstream picture was murkier, and without product data, we were only ever seeing half the story.

That gap between acquisition and retention is where most marketing strategies quietly fail. You can spend your way to impressive traffic numbers and still be running a business with a leaky bucket. Product analytics is what tells you the bucket is leaking.

The connection between marketing and product data is also where BCG’s research on data-driven organisations lands. Businesses that connect behavioural data across the customer lifecycle consistently outperform those that optimise in siloes. Marketing optimising for clicks while product optimises for engagement, with no shared data layer between them, is a structural inefficiency that compounds over time.

The Core Metrics That Actually Matter

Product analytics generates a lot of data. The discipline is in knowing which metrics to prioritise and which to ignore. These are the ones that tend to have the most commercial weight.

Activation Rate

Activation measures whether a new user reaches the moment where they experience the core value of your product. This varies by product. For a project management tool it might be creating a first project and inviting a collaborator. For an e-commerce app it might be completing a first purchase. The activation event is the threshold between someone who signed up and someone who got the point.

Low activation rates are almost always a product or onboarding problem, not a marketing problem. But marketing teams need to know the activation rate because it directly affects the real cost of acquisition. If you’re paying to acquire 1,000 users and only 200 activate, your true cost per activated user is five times your headline CPA. That changes the economics of every channel decision you make.

Retention Curves

Retention is the metric that separates products that work from products that don’t. A retention curve shows what percentage of users who joined in a given period are still active at day 7, day 30, day 90, and beyond. A healthy retention curve flattens out at some point, indicating a core group of users who have made the product part of their routine. A curve that keeps declining toward zero indicates a product that isn’t sticky enough to sustain itself on acquisition alone.

For marketers, the shape of the retention curve should inform acquisition strategy directly. Pouring budget into a product with a flat-to-zero retention curve is expensive and in the end futile. I’ve seen this play out more than once in agency reviews, where a client was scaling paid acquisition aggressively on a product with fundamental retention problems. The channel metrics looked fine. The business was bleeding.

Feature Adoption

Feature adoption tells you which parts of your product users actually engage with. This matters for marketing because it tells you what to talk about. If a feature that the product team considers core is only being used by 15% of users, that’s either a communication problem or a design problem. Marketing can address the communication angle. Product needs to address the design angle. But you can’t have that conversation without the data.

Funnel Completion and Drop-off

Every product has flows: onboarding, checkout, setup, upgrade. Funnel analysis shows where users drop out of these flows and at what rate. This is where product analytics and conversion rate optimisation overlap most directly. Integrating behavioural data with A/B testing lets you move from identifying where users drop off to testing hypotheses about why, and validating fixes before rolling them out broadly.

Session Frequency and Depth

How often do users return, and how much do they do when they’re there? These metrics are proxies for engagement quality. A user who logs in once a week and completes multiple actions is more valuable than one who logs in daily but bounces immediately. Segmenting users by session frequency and depth often reveals distinct behavioural cohorts that should be treated differently in both product development and marketing communications.

Where Event Tracking Goes Wrong

The quality of your product analytics is entirely dependent on the quality of your event tracking. This is where most implementations fall apart, not because the tools are bad, but because the event taxonomy wasn’t designed carefully enough before implementation began.

Common problems include: events named inconsistently across platforms, making cross-device analysis unreliable; events that fire on page load rather than on actual user action, inflating engagement numbers; and event names that made sense to the developer who wrote them but are meaningless to anyone doing analysis six months later.

I’ve sat in analytics reviews where the team was confidently reporting on “button_click” as a key engagement metric, with no specification of which button, on which page, in which context. The number was meaningless. But it was in the dashboard, so it got reported. This is the kind of thing that happens when tracking is implemented without a measurement plan, and it’s more common than most teams would admit.

The fix is straightforward in principle, though it requires discipline in practice. Define your event taxonomy before you build it. Name events in a way that describes the user action and its context. Document what each event means and what properties it should carry. Treat your tracking plan as a living document that gets updated when the product changes. Failing to prepare in analytics is preparing to fail, and that’s as true for product event tracking as it is for any other measurement framework.

UTM parameters are the bridge between your marketing data and your product data. If your campaigns aren’t properly tagged, you can’t connect acquisition source to downstream product behaviour. Getting UTM tracking right is a prerequisite for any meaningful analysis of how different marketing channels produce different quality users.

Connecting Marketing Channels to Product Behaviour

One of the most valuable things product analytics enables is a proper comparison of user quality across acquisition channels. Not all traffic is equal, and the channel that drives the most sign-ups is rarely the channel that drives the most retained, high-value users.

When I was growing teams at iProspect, we were managing substantial paid media budgets across multiple channels for large clients. The channel-level metrics were strong. But the clients who were most commercially sophisticated were the ones who pushed us to connect channel performance to downstream outcomes: repeat purchase rates, lifetime value, product engagement. The ones who just optimised for CPA were often optimising for the wrong thing.

The analysis you want to run is a cohort comparison by acquisition source. Take users acquired from paid search in a given month. Take users acquired from organic in the same period. Take users acquired from email or referral. Then compare their activation rates, 30-day retention, and revenue contribution over 90 days. This analysis almost always produces surprises. Channels that look expensive on a CPA basis often produce users with significantly higher retention. Channels that look efficient often produce users who churn quickly.

This kind of analysis requires your marketing data and your product data to be in the same place, or at least queryable together. That’s a data infrastructure question as much as an analytics question, but it’s the work that separates teams doing real measurement from teams doing reporting theatre.

Tools like Hotjar sit alongside product analytics platforms to add qualitative depth to quantitative signals. Combining session recording and heatmap data with behavioural analytics helps explain why users behave the way the numbers describe, which is often the more useful question.

The Limits of Product Analytics Data

Product analytics tools, like all analytics tools, are a perspective on reality rather than reality itself. It’s worth being clear about the gaps.

Consent and privacy regulations mean that a meaningful proportion of users will decline tracking, particularly in markets with active cookie consent enforcement. Your product analytics data represents opted-in users, who may behave differently from users who declined. This doesn’t make the data useless, but it does mean you’re working with a sample, not a census.

Ad blockers and browser privacy features suppress some tracking, particularly for JavaScript-based event tracking. The scale of this varies by audience. A developer-focused product with a technically literate user base will have significantly higher tracking suppression than a consumer product. The reasons analytics data is inaccurate are well documented, and product analytics tools face the same structural challenges as web analytics platforms.

There’s also the question of what the data doesn’t capture. Product analytics tells you what users did. It doesn’t tell you why. It doesn’t capture the conversation a user had with a colleague before deciding to upgrade. It doesn’t capture the frustration that led someone to churn without ever triggering a specific event. Quantitative product data needs qualitative research alongside it, user interviews, support ticket analysis, churn surveys, to be properly interpreted.

I’ve seen teams make expensive product decisions based on funnel drop-off data that turned out to be a tracking bug rather than a genuine user behaviour pattern. The event was firing at the wrong point in the flow, making a step look like a drop-off when users were actually completing it. The data looked clean. It wasn’t. Validation of your tracking implementation isn’t a one-time task. It needs to happen regularly, especially after product changes.

Building a Product Analytics Practice That Actually Works

The gap between having a product analytics tool and having a product analytics practice is significant. Most teams have the tool. Fewer have the practice.

A practice means: questions get asked before dashboards get built. It means the metrics in the dashboard connect to decisions someone is actually making. It means the team reviews the data regularly and acts on what it shows. It means the tracking implementation is maintained as the product evolves, rather than left to decay as features change and events stop firing correctly.

Start with the decisions you need to make, not the data you could collect. The most useful framing is: what would change our behaviour if we knew it? If the answer is “nothing”, you don’t need that metric. If you’re collecting data that nobody acts on, you’re adding complexity without value. The power of analytics comes from focus, not from comprehensiveness.

Define your north star metric early. This is the single metric that best captures whether your product is delivering value to users. For a communication tool it might be messages sent. For a content platform it might be content pieces consumed per week. For a B2B SaaS product it might be the number of active users per account. Everything else in your analytics framework should connect back to this metric in some way.

Then build your funnel from acquisition through to the behaviours that predict retention. Map the events that represent progress through that funnel. Instrument them carefully. Review them regularly. When the product changes, update the tracking plan before the feature ships, not after.

Early in my career, when I couldn’t get budget for a proper website build, I taught myself to code and built it myself. The lesson wasn’t really about coding. It was about not waiting for perfect conditions before getting something useful done. Product analytics is similar. You don’t need the most sophisticated tool or a perfect data infrastructure to start getting useful answers. You need clear questions, clean tracking, and the discipline to act on what you find.

If you’re working through the broader measurement challenge, including attribution, GA4 configuration, and how to connect channel data to business outcomes, the Marketing Analytics section of The Marketing Juice covers the full landscape with the same focus on commercial usefulness over analytical theatre.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions