Digital Marketing Analytics: Stop Measuring Everything, Start Measuring What Matters

Digital marketing analytics is the practice of collecting, measuring, and interpreting data from your online marketing activity to understand what is driving business outcomes and what is not. Done well, it gives you a clear line of sight between marketing spend and commercial results. Done badly, it produces dashboards full of numbers that make everyone feel busy while the business bleeds budget.

Most marketing teams sit closer to the second version than they would admit.

I have been working in and around marketing analytics since roughly 2000, across agencies, client-side roles, and everything in between. I have managed hundreds of millions in ad spend across more than 30 industries. I have sat in rooms where the data said one thing and the business results said something entirely different. The gap between those two things is where most analytics problems live, and it is almost never a technology problem.

Why Most Analytics Setups Produce Noise, Not Insight

The first thing I notice when I look at a new client’s analytics setup is not what they are measuring. It is what they think the numbers mean. Those are rarely the same thing.

Early in my career, I asked the managing director of the agency I was working at for budget to build a new website. The answer was no. So I taught myself to code and built it myself. That experience gave me something most marketers never get: a genuine understanding of how web data is generated at the source. When you have written the tracking code yourself, you know exactly how fragile it is. You know that a misplaced tag, a misconfigured event, or a browser that blocks cookies can quietly corrupt weeks of data without triggering a single alert. Most people reading dashboards have no idea how the numbers got there. That gap in understanding is where bad decisions are made.

The analytics industry has spent twenty years selling the idea that more data equals better decisions. It does not. It equals more time spent in reporting meetings arguing about numbers instead of acting on them. MarketingProfs identified this problem over a decade ago, noting that web analytics only delivers value when marketers know what questions they are trying to answer before they look at the data. That observation has aged remarkably well.

If you want a broader grounding in how analytics fits into the wider marketing measurement landscape, the Marketing Analytics hub covers the full picture, from GA4 setup to attribution modelling to ROI measurement.

What Digital Marketing Analytics Actually Measures

It is worth being precise about this, because the term covers a lot of ground and the conflation of different measurement types causes real problems in practice.

At its most basic, digital marketing analytics measures three things: what people do on your digital properties, where they came from, and what happened as a result. Everything else is a variation or a combination of those three things.

Traffic analytics tells you about volume and source. Engagement analytics tells you about behaviour on site or in-app. Conversion analytics tells you about outcomes. The mistake most teams make is treating these as a hierarchy, where conversion is the only metric that matters, and everything else is context. That is partially right but not entirely. Engagement data, interpreted correctly, tells you things about intent and friction that conversion data alone cannot. Time on page, for example, is a more nuanced signal than it appears, particularly when you understand how GA4 calculates it differently from Universal Analytics.

There is also a category of analytics that most teams underinvest in: competitive and market intelligence. Knowing what your own campaigns are doing is necessary but not sufficient. Knowing how that performance compares to market conditions, seasonal trends, and competitor activity is what separates reactive reporting from genuine strategic insight.

One area worth calling out specifically is the measurement of newer channel types. As formats evolve, so does the complexity of attribution. Understanding how to measure the effectiveness of AI avatars in marketing is a good example of how measurement frameworks need to adapt as channels do. The principles are consistent even when the mechanics are new.

The Attribution Problem Nobody Wants to Talk About Honestly

Attribution is the single most contested area in digital marketing analytics, and the industry’s handling of it has been, to put it charitably, inconsistent.

The core problem is straightforward. A customer interacts with your brand multiple times before converting. Each of those interactions happened in a different channel, possibly on a different device, possibly days or weeks apart. Your analytics platform has to decide which of those interactions gets credit for the conversion. Every decision it makes is a simplification. Every simplification introduces distortion. The question is not whether your attribution model is accurate. It is not. The question is whether it is useful, and whether you understand its limitations well enough to make good decisions with it.

I have judged the Effie Awards, which means I have seen how the best-performing marketing in the world gets measured and reported. One thing that consistently separates the strongest entries from the weaker ones is not the sophistication of their analytics stack. It is the honesty with which they acknowledge what their measurement can and cannot tell them. The teams that win are the ones who build a coherent argument from imperfect data, not the ones who pretend the data is perfect.

Forrester has been making this point for years. Their work on marketing measurement snake oil is worth reading if you have not already, particularly their observation that vendors routinely oversell the precision of their attribution models in ways that do not hold up under scrutiny.

For a deeper treatment of how attribution models work and where they break down, the article on attribution theory in marketing covers the conceptual foundations in detail. Understanding the theory is what allows you to interrogate the outputs rather than just accepting them.

There is also a specific version of this problem that affects affiliate and partnership channels. The question of how to measure affiliate marketing incrementality gets to the heart of what attribution is actually trying to answer: not just which channel got credit, but which channel genuinely caused the conversion. Those are very different questions.

Vanity Metrics and the Organisations That Love Them

Every organisation I have ever worked with has had at least one vanity metric embedded in its reporting. Usually more than one. And usually, the vanity metric has been there long enough that nobody remembers why it was included in the first place.

Vanity metrics are not just useless. They are actively harmful, because they create the appearance of performance where none exists. When a team is hitting its targets on impressions, reach, and follower growth while the business is not growing, the metrics are not failing to capture reality. They are obscuring it. That is a different and more serious problem.

The most common vanity metrics I encounter in client reporting are: total website sessions without segmentation, social media reach without engagement or conversion context, email open rates without click-through or revenue correlation, and cost per click without any connection to downstream value. None of these numbers is inherently meaningless. All of them become meaningless when reported in isolation, without the connective tissue that links them to business outcomes.

When I was growing an agency from 20 to 100 people, one of the disciplines I had to install early was a distinction between activity metrics and outcome metrics. Activity metrics tell you what the team is doing. Outcome metrics tell you whether it is working. Both matter, but they answer different questions and should never be conflated in the same reporting conversation. The moment you start using activity metrics to justify outcome targets, you have a problem that no amount of better data will fix.

Building a Measurement Framework That Actually Works

A measurement framework is not a list of KPIs. It is a structured set of questions about your business, mapped to the data sources and metrics that can answer them, with a clear understanding of what each metric can and cannot tell you.

That definition matters because most “measurement frameworks” I encounter are actually just KPI lists dressed up with a bit of structure. They tell you what to measure but not why, and they do not acknowledge the limitations of the measurement. That is not a framework. It is a reporting template.

A genuine measurement framework starts with business questions, not metrics. What do we need to know to make better decisions? What decisions are we actually trying to make? Which of those decisions are being made with inadequate information right now? Those questions, answered honestly, will tell you far more about what to measure than any analytics best-practice checklist.

Preparation in web analytics is not optional. The teams that define their measurement approach before a campaign launches consistently extract more useful insight from the same data than teams that try to retrofit measurement after the fact. This is not a complicated observation, but it is one that a surprising number of teams ignore.

The practical steps look like this. First, define the business outcome you are trying to drive. Second, identify the leading indicators that predict that outcome. Third, map those indicators to the data sources that can measure them. Fourth, acknowledge explicitly what your measurement cannot capture. Fifth, build your reporting around decisions, not data points.

A well-structured marketing dashboard is one output of this process, but it is not the process itself. The dashboard is where the framework becomes visible. The thinking happens before you open any reporting tool.

GA4, Platform Analytics, and the Limits of Any Single Tool

Google Analytics 4 is the dominant web analytics platform for most marketing teams, and it is genuinely more capable than Universal Analytics in several important ways. The event-based data model is more flexible. The cross-device tracking is more sophisticated. The integration with Google’s advertising ecosystem is tighter. These are real improvements.

But GA4 is still a tool with significant blind spots, and the transition from Universal Analytics has introduced new ones. The data that Google Analytics goals cannot track is a useful starting point for understanding where the gaps are. Phone calls, offline conversions, cross-device journeys that break the tracking chain, and behaviour in environments where cookies are blocked or restricted are all areas where GA4’s picture is incomplete. Knowing what you cannot see is as important as knowing what you can.

There is also a broader question about tool dependency. When a single platform is both your primary analytics tool and a major advertising vendor, there is an inherent conflict of interest in how that platform reports on its own performance. I am not suggesting the data is fabricated. I am suggesting that the framing of that data, the default attribution windows, the conversion counting methodology, and the way cross-channel comparisons are presented, all reflect choices that are not neutral. Moz’s overview of Google Analytics alternatives is worth reading not because you should necessarily switch, but because understanding what other tools measure differently will sharpen your thinking about what GA4 is actually doing.

Forrester’s caution about black box analytics is directly relevant here. When you cannot see how a number was calculated, you cannot evaluate whether it is trustworthy. That applies to GA4’s data-driven attribution model just as much as it applies to any other algorithmic output.

A/B testing within GA4 is one area where the platform’s capabilities are genuinely useful. Semrush’s guide to A/B testing in GA4 covers the mechanics well. The principle worth emphasising is that testing is only valuable when you have a clear hypothesis before you start. Testing without a hypothesis is just collecting data. It is not learning.

Measuring Channels That Resist Easy Measurement

Some of the most commercially important marketing activity is also the hardest to measure. Brand activity, content marketing, and organic search all have long feedback loops and attribution challenges that performance marketers often use as an excuse to deprioritise them in favour of channels where the numbers look cleaner.

That trade-off is usually a mistake. The channels that are easiest to measure are often the ones that are capturing demand that already exists, rather than creating new demand. When I was at lastminute.com, I launched a paid search campaign for a music festival and saw six figures of revenue within roughly a day. It felt like a brilliant campaign. Looking back, a significant portion of that revenue was from people who were already going to buy tickets and just happened to click a paid link rather than an organic one. The measurement made it look better than it was, because the measurement could not see the counterfactual.

Measuring inbound marketing, which operates on longer timescales and through less direct conversion paths, requires a different approach. The article on inbound marketing ROI covers how to build a case for channels where the attribution is genuinely difficult. The short version is that you need a combination of assisted conversion data, cohort analysis, and honest acknowledgment of what you cannot directly attribute.

The same principle applies to emerging formats. Understanding how to measure the success of generative engine optimisation campaigns is a current example of a channel where the measurement infrastructure is still catching up to the activity. The instinct to wait until measurement is perfect before investing is understandable but commercially dangerous. Better to invest with honest approximation than to ignore a channel because the numbers are messy.

The Organisational Problem Behind Most Analytics Failures

Most analytics failures are not technology failures. They are organisational failures. The data was available. The tools were capable. The insight was there if someone had looked for it. But the organisation was not structured to act on what the data was saying.

I have seen this pattern repeatedly across agency and client-side roles. The analytics team produces reports. The marketing team reads the headlines and ignores the detail. The commercial team does not trust the numbers. The board asks for a metric that nobody defined properly. And the whole cycle repeats without anyone stopping to ask whether the measurement setup is actually serving the decisions that need to be made.

The fix is not a better analytics platform. It is a clearer conversation between the people who generate the data and the people who are supposed to use it to make decisions. That conversation needs to happen before the campaign launches, not after the results come in.

Analytics is not a reporting function. It is a decision-support function. When it is positioned as the former, it produces dashboards. When it is positioned as the latter, it produces better marketing decisions. That distinction, more than any technical capability, is what separates organisations that get genuine value from their analytics investment from those that just have very detailed reports of things that already happened.

If you are building or rebuilding your analytics capability, the Marketing Analytics hub is where I have collected the most practically useful material on measurement, attribution, and making sense of GA4. It is worth bookmarking as a reference rather than treating it as a one-time read, because the landscape shifts and the frameworks need to shift with it.

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