Data Analytics Strategy: Stop Measuring Everything, Start Measuring What Matters

A data analytics strategy is a structured approach to deciding what you measure, why you measure it, and what decisions that measurement is supposed to inform. Most teams skip this and go straight to the tools, which is why they end up with dashboards full of numbers that nobody acts on.

The problem is rarely a shortage of data. It is a shortage of clarity about what the data is for.

Why Most Analytics Strategies Fail Before They Start

I have sat in a lot of analytics reviews across a lot of organisations. The pattern is almost always the same. Someone pulls up a dashboard showing sessions, bounce rate, time on page, and a handful of conversion metrics. The room nods. Someone asks why organic traffic dropped in February. Nobody knows. The meeting ends with an action to “look into it.”

Nothing changes, because the measurement was never connected to a decision in the first place.

When I was growing an agency from around 20 people to over 100, one of the things that became clear early on was that the teams doing the best analytical work were not the ones with the most sophisticated tooling. They were the ones who could answer a simple question before opening any platform: what decision does this data need to support? If you cannot answer that, you are doing reporting, not analytics.

The distinction matters. Reporting tells you what happened. Analytics tells you why it happened and what you should do differently. Most organisations are very good at the first and poor at the second, because the second requires strategic thinking and the first just requires a scheduled export.

If you want to go deeper on the measurement infrastructure that underpins all of this, the Marketing Analytics hub covers GA4 setup, attribution, and the broader measurement stack in detail.

What a Data Analytics Strategy Actually Consists Of

Strip away the frameworks and the consulting language, and a data analytics strategy has four components. You need to know what you are trying to achieve commercially. You need to know which metrics are genuinely connected to those outcomes. You need a reliable way to collect and access that data. And you need a process for turning the data into decisions.

Most teams have the third component. Some have parts of the second. Very few have the first and fourth in a form that actually governs how they work.

Start with the commercial objectives. Not marketing objectives. Commercial ones. Revenue, margin, customer acquisition cost, lifetime value, retention rate. These are the numbers the business cares about. Your analytics strategy should trace a clear line from those numbers back to the marketing activity that influences them.

That line is rarely as clean as people want it to be. A paid search campaign might drive form fills, which become qualified leads, which convert to customers at a certain rate, which retain at another rate, which produce a lifetime value that either justifies the acquisition cost or does not. If your analytics only measures the form fills, you are missing most of the picture. Unbounce has written sensibly about simplifying analytics without losing commercial relevance, and the principle holds: simpler is only better if it is still connected to outcomes that matter.

The Vanity Metric Problem Is Worse Than You Think

Vanity metrics are not just unhelpful. They are actively harmful, because they create the appearance of progress where none exists and they compete for attention with the metrics that actually matter.

I judged the Effie Awards for a period, which gives you an unusual vantage point on marketing effectiveness. The entries that struggled were rarely the ones with bad creative. They were the ones where the success metrics had drifted away from business outcomes. Campaigns that generated enormous reach and engagement but could not demonstrate commercial impact. The measurement framework had been built to make the campaign look good rather than to assess whether it worked.

That is a strategic failure, not a measurement failure. The tools were working fine. The question being asked of them was wrong.

Social media reach is the classic example. Page views is another. Time on site is a third. None of these are inherently useless, but all of them become useless the moment they are treated as proxies for commercial value without evidence that the relationship exists. If you can demonstrate that longer time on site correlates with higher purchase intent in your specific context, then it is worth tracking. If you cannot, it is decoration.

The MarketingProfs piece on web analytics makes a point that has aged well: the power of analytics is not in the volume of data you collect, it is in the quality of the questions you bring to it. That was true when it was written and it is more true now, when the volume of available data has increased by an order of magnitude.

Building Your Measurement Framework Backwards

The right way to build a measurement framework is to start at the end and work backwards. Begin with the commercial outcome you are trying to influence. Then identify the leading indicators that predict movement toward that outcome. Then identify the marketing activities that drive those leading indicators. Then, and only then, decide what to measure.

In practice, this looks like the following. If the commercial outcome is new customer revenue, the leading indicator might be qualified pipeline. The marketing activity driving that pipeline might be content-driven organic search. The metrics worth tracking are therefore organic sessions from high-intent queries, conversion rate from those sessions to qualified leads, and the downstream conversion from lead to customer. Everything else is context at best and noise at worst.

This is not complicated. But it requires discipline, because the instinct in most organisations is to measure what is easy to measure rather than what is important to measure. GA4 makes it trivially easy to pull sessions, pageviews, and engagement rate. It is considerably harder to connect those numbers to revenue, which is exactly why most teams stop at the easy numbers.

Early in my career, when I was teaching myself to build websites because the budget for a developer did not exist, I learned something that has stayed with me ever since. When you have limited resources, you become very focused on what actually matters. You do not have the luxury of measuring everything. You measure what tells you whether you are moving in the right direction. That constraint turns out to be a useful discipline even when resources are not limited.

The MarketingProfs piece on analytics preparation frames this well: the failure mode in web analytics is almost always a failure of preparation, not a failure of technology. Teams that invest in defining what they want to measure before they configure their tools get dramatically more useful output than teams that configure first and think later.

What GA4 Can and Cannot Tell You

GA4 is the default analytics tool for most marketing teams, and it is genuinely capable when it is set up well. The problem is that most implementations are not set up well, and even the best implementations have significant limitations that teams tend to underestimate.

The data model in GA4 is event-based, which is more flexible than Universal Analytics but also more demanding to configure correctly. Out of the box, you get a reasonable picture of traffic and basic engagement. To get anything commercially useful, you need to define your own events, set up conversion tracking, and connect the data to whatever CRM or revenue system holds your actual business outcomes. Most teams do some of this. Very few do all of it.

Semrush’s overview of GA4 is a useful reference for the setup fundamentals. Crazyegg’s guide to GA4 filters covers one of the most commonly overlooked configuration steps, which is filtering out internal traffic and bot traffic before you draw any conclusions from your data. If you have not done this, a meaningful proportion of what you are looking at is not real user behaviour.

Beyond configuration, there is a more fundamental limitation. GA4 uses modelled data to fill gaps created by consent refusals, ad blockers, and cross-device behaviour. The modelling is generally reasonable, but it means the numbers you see are an estimate, not a count. The margin of error varies by site, by audience, and by traffic mix. For most B2C sites with broad audiences, the modelled data is probably close enough to be directionally useful. For B2B sites with smaller, more technically sophisticated audiences, the gap between reported and actual behaviour can be significant.

This is not a criticism of GA4. It is a structural reality of web analytics in a privacy-first environment. The appropriate response is not to distrust the data entirely, but to treat it as one input among several rather than as ground truth. For teams that need more granular or reliable data, exporting GA4 data to BigQuery opens up analytical possibilities that the standard interface does not support, including unsampled analysis and more sophisticated attribution modelling.

The Dashboard Problem

Most marketing dashboards are built to be comprehensive. They should be built to be decisive.

A comprehensive dashboard shows everything. A decisive dashboard shows the metrics that, when they move, require a response. Those are very different things, and the difference has real consequences for how teams spend their time.

I have seen agencies spend hours every week producing reporting packs that clients glance at and file. The problem is not that clients are uninterested in data. It is that the data is not connected to anything they need to decide. When the data does not drive decisions, it becomes a ritual rather than a tool.

Crazyegg’s guide to building GA4 dashboards makes a practical point about starting with the questions you need answered rather than the metrics available to you. That is the right starting point. For each metric on your dashboard, you should be able to answer: what decision does this inform, and what would we do differently if this number moved significantly in either direction? If you cannot answer that, the metric does not belong on the dashboard.

The discipline of building a decisive dashboard is also a useful diagnostic for your analytics strategy overall. If you find it difficult to identify the metrics that require a response when they move, it usually means your measurement framework is not tightly connected to your commercial objectives. The dashboard problem is often a strategy problem in disguise.

Connecting Digital Analytics to Revenue

The gap between digital analytics and revenue is where most analytics strategies break down. You can measure sessions, conversions, and cost per acquisition with reasonable confidence. Connecting those numbers to actual revenue, margin, and customer lifetime value requires data that usually lives outside your analytics platform.

The practical solution is data integration. Your CRM holds lead quality and conversion rates. Your finance system holds revenue and margin. Your customer success data holds retention and lifetime value. None of these are in GA4 by default, but all of them can be connected to it with varying degrees of effort.

When I was managing significant paid media budgets across multiple clients, the single most valuable analytical improvement we made was not a new tool or a new attribution model. It was connecting the paid search conversion data to the CRM so we could see which campaigns were generating revenue rather than just leads. The shift in budget allocation that followed was substantial, because the campaigns that looked best on cost per lead looked very different when you could see the downstream conversion rates and deal values.

For teams using video as part of their content mix, Wistia’s GA4 integration is worth knowing about. Video engagement data sitting in a separate platform is useful. Video engagement data connected to your broader analytics stack, so you can see how it correlates with conversion behaviour, is considerably more useful.

The broader principle is that analytics value compounds when data sources are connected. A session in GA4 that you can trace through to a closed deal in your CRM is worth far more analytically than a session you can only see as a pageview. The technical work to make those connections is not trivial, but the commercial clarity it creates justifies the investment.

How to Prioritise What to Measure

If your team is starting from a position of measuring too much without enough clarity, the practical question is how to prioritise. You cannot fix everything at once, and attempting to do so usually results in fixing nothing.

Start with the commercial decisions that matter most in the next 90 days. Not the metrics you track. The decisions you need to make. Where should budget go? Which channels should we scale? Which campaigns should we cut? Which audience segments are worth investing in? Work backwards from those decisions to identify the data gaps that are making them harder than they need to be.

That exercise will almost always surface two or three specific measurement problems that, if solved, would materially improve decision quality. Fix those first. Then repeat the exercise.

This is a more useful approach than attempting to build a comprehensive analytics strategy from scratch, because it connects measurement improvement to commercial value from the start. Every measurement problem you solve should make a specific decision easier or more reliable. If it does not, it is probably not the right problem to solve first.

One thing worth being honest about: analytics strategy is never finished. The business objectives change, the channels change, the tools change, and the data landscape changes. The goal is not a perfect measurement framework. It is a framework that is good enough to support the decisions you need to make now, with a process for improving it as those decisions evolve.

That mindset, honest approximation rather than false precision, is what separates teams that use analytics well from teams that use analytics to feel productive. The numbers are not the point. The decisions are the point. The numbers just need to be reliable enough to make those decisions better than they would be without them.

There is more on building the right measurement infrastructure across channels and tools in the Marketing Analytics section of The Marketing Juice, including pieces on GA4 configuration, attribution models, and how to make sense of data that does not always tell a consistent story.

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