Marketing Analysis Framework: Stop Measuring Activity, Start Measuring Causality

A marketing analysis framework is a structured approach to evaluating what your marketing is actually doing for the business, not just what it appears to be doing. Done well, it separates signal from noise, connects marketing activity to commercial outcomes, and gives leadership a clear picture of where growth is coming from and why.

Done badly, it becomes an elaborate exercise in confirmation bias, a dashboard full of metrics that make the team feel productive without telling anyone whether the marketing is working.

Most frameworks sit closer to the second description than the first.

Why Most Marketing Analysis Gets the Wrong Answer

Early in my career, I was a true believer in performance marketing. We had dashboards, attribution models, cost-per-acquisition figures that looked clean and defensible. When a client asked whether the marketing was working, I could point to numbers. It felt rigorous.

It took me years to fully accept how much of what we were measuring was not causation. We were tracking people who were already going to buy and counting the last touchpoint before purchase as the thing that made it happen. The attribution model said paid search was driving growth. The reality, in many cases, was that paid search was sitting at the bottom of a funnel that the brand, the product, and the customer experience had already filled.

I think about it like a clothes shop. If someone tries something on, they are ten times more likely to buy it. But the act of trying it on did not create the desire to visit the shop. Something else did that. If you only measure conversions at the till, you will never understand what actually drove the footfall. You will just keep optimising the fitting rooms.

This is the foundational problem with most marketing analysis frameworks. They are built around the data that is easiest to collect, which tends to be lower-funnel, short-term, and heavily skewed toward channels that can attach themselves to a conversion event. The harder questions, about awareness, brand preference, new audience reach, and long-run demand creation, get treated as unmeasurable and therefore unimportant.

They are not unmeasurable. They are just harder to measure. And harder to measure does not mean less valuable.

The Four Layers of a Sound Marketing Analysis Framework

A framework that actually serves the business needs to operate at four levels. Most teams only run one of them. Here is how I think about the structure.

Layer 1: Market Context

Before you look at a single campaign metric, you need to understand what is happening in the market. Is the category growing or contracting? Is demand seasonal? Are there macro conditions, economic, competitive, regulatory, that are moving the numbers independently of anything your marketing did?

I have sat in too many quarterly reviews where a team celebrated a 20% uplift in leads without anyone asking whether the whole market had moved 25% in the same direction. If it had, the marketing had actually underperformed. Context makes the difference between a correct interpretation and a flattering one.

Tools like Google Trends, category search volume data, and competitor share-of-voice tracking give you a rough read on market movement. They are imperfect. But even an imperfect market context layer is better than no context at all. Understanding market penetration dynamics is a useful starting point for framing what realistic growth looks like in a given category.

Layer 2: Audience Behaviour

The second layer asks who is actually engaging with your marketing, and whether that audience matches the people you need to reach to grow the business.

This is where most analysis frameworks have a significant blind spot. They measure volume, clicks, impressions, open rates, but not composition. You can have a highly efficient campaign that is almost entirely reaching existing customers or people already deep in the consideration phase. That campaign will look excellent on a cost-per-conversion basis. It will do almost nothing to extend your reach into genuinely new audiences.

When I was growing an agency from around 20 people to over 100, one of the things I learned was that new business did not come from the same channels as retention. The activities that kept existing clients happy were completely different from the activities that introduced us to new ones. Conflating the two in a single performance report would have given us a dangerously misleading picture of what was driving growth.

Audience analysis tools, including on-site behaviour platforms and CRM segmentation, help here. Hotjar and similar tools give you a qualitative layer on top of the quantitative data, showing not just that people visited, but what they did and where they dropped off. That behavioural texture matters when you are trying to understand whether your marketing is reaching and converting the right people.

Layer 3: Channel Performance

This is the layer most teams are actually running. Impressions, clicks, cost-per-click, conversion rate, return on ad spend, cost per lead. These metrics are useful. They are also deeply insufficient on their own.

The problem is not the metrics themselves. It is that they are treated as the whole picture rather than one layer of it. When channel performance is the only thing being analysed, the business ends up optimising for the metrics rather than for growth. Campaigns get refined to death. Budgets flow to the channels with the cleanest attribution. The brand slowly starves.

I have managed hundreds of millions in ad spend across more than thirty industries. The pattern I see repeatedly is that teams with sophisticated channel-level reporting often have the least accurate picture of what is actually driving their business. The sophistication of the measurement creates false confidence. You feel like you know what is happening because the dashboard is detailed. But a detailed view of one layer is not the same as understanding the system.

Channel performance analysis should include incrementality thinking, even if you cannot run a formal incrementality test. Ask: if we turned this channel off, what would actually happen to sales? The honest answer to that question is often more instructive than six months of attribution data.

Layer 4: Commercial Outcome

The final layer connects marketing activity to the numbers that actually matter to the business: revenue, margin, customer lifetime value, market share, and net new customer acquisition.

This is where most marketing analysis frameworks fall apart entirely. Not because marketers do not care about commercial outcomes, but because the connection between marketing activity and those outcomes is genuinely hard to establish with confidence. So the analysis stops at channel performance, and the commercial layer gets filled in with assumptions.

The honest approach is to acknowledge the uncertainty and work with it rather than paper over it. You may not be able to prove with precision that a brand campaign drove a 3% lift in revenue. But you can track revenue trends over time, control for market conditions, and build a reasonable case for whether the marketing is contributing to commercial health or not.

Judging the Effie Awards gave me an interesting window into how the best marketing teams in the world make this connection. The entries that stood out were not the ones with the most sophisticated attribution models. They were the ones that could tell a coherent story from marketing activity through to business result, with honest acknowledgement of what they could and could not prove.

If you want to go deeper on how this connects to broader growth strategy, the Go-To-Market and Growth Strategy hub covers the commercial context that should sit behind any analysis framework.

The Incrementality Problem Nobody Wants to Talk About

There is a conversation that does not happen often enough in marketing reviews, and it goes something like this: how much of what we are attributing to our marketing would have happened anyway?

It is an uncomfortable question because the honest answer is often “quite a lot.” Branded search captures people who were already looking for you. Retargeting reaches people who had already visited your site. Email converts people who had already signed up. These are not worthless activities. But they are not creating demand. They are harvesting it.

A sound marketing analysis framework has to grapple with this. Forrester’s thinking on intelligent growth has long argued that sustainable growth requires reaching beyond existing intent, not just optimising conversion of people already in market. That framing holds up.

The practical implication for analysis is that you need to track metrics that are leading indicators of demand creation, not just demand capture. Brand search volume trends. Share of voice in the category. New-to-brand customer ratios. Unprompted awareness scores in target segments. These are harder to track and slower to move. They are also closer to the actual levers of long-run growth.

I have worked with businesses that had excellent conversion metrics and declining market share simultaneously. The marketing looked efficient. The business was quietly losing ground. The analysis framework was not built to catch that, because it was only looking at the bottom of the funnel.

What Good Analysis Actually Looks Like in Practice

A practical marketing analysis framework does not need to be complicated. It needs to be honest. Here is the structure I would use.

Start with a monthly market context review. Fifteen minutes looking at category search trends, competitor activity, and any relevant external conditions. This takes almost no time and prevents a lot of misinterpretation.

Run a quarterly audience composition audit. Look at who is actually engaging with your marketing. What proportion are new to brand? What is the demographic and behavioural profile of converters versus browsers? Are you reaching the audiences you need to reach to grow, or are you increasingly efficient at talking to the same people?

Keep channel performance reporting on a regular cadence, but discipline yourself to ask the incrementality question for each channel at least once a quarter. Not “what did this channel deliver?” but “what would have happened without it?”

Build a simple commercial health dashboard that sits above the channel metrics. Revenue trends, new customer acquisition rate, customer lifetime value trajectory, and market share where you can track it. Review this monthly alongside the channel data, not separately from it.

And finally, make the analysis honest. The most dangerous thing a marketing team can do is build a framework that is designed to confirm that the marketing is working. I have seen this happen in agencies and in-house teams alike. The metrics get selected because they tell a good story. The story gets told to leadership. Leadership allocates more budget. The underlying business problem, whether it is product, pricing, or customer experience, goes unaddressed because the marketing dashboard looks healthy.

Marketing is sometimes a blunt instrument used to prop up businesses with more fundamental issues. A good analysis framework will surface that, not hide it. BCG’s research on go-to-market strategy makes a related point: sustainable commercial performance requires alignment between marketing, product, and the broader business, not just optimised campaign execution.

Common Mistakes That Undermine Marketing Analysis

A few patterns come up repeatedly in teams that are struggling to get useful insight from their analysis.

The first is mistaking correlation for causation. Sales went up in the quarter when we ran the campaign, therefore the campaign drove the sales. This is one of the most persistent errors in marketing analysis, and it is especially dangerous when it is used to justify large budget commitments. Correlation is a starting point for investigation, not a conclusion.

The second is over-relying on last-click attribution. Most attribution models in common use are not measuring what drove a purchase decision. They are measuring what happened to be the last recorded touchpoint before a conversion event. These are very different things. Even growth-focused practitioners acknowledge that attribution models are a useful approximation, not a precise account of how decisions are made.

The third is analysing channels in isolation. Paid social, paid search, email, and content do not operate independently. They interact. A brand awareness campaign on social may lift branded search volume two weeks later. If you are only looking at paid search performance, you will not see that connection. You may even cut the social campaign because it looks expensive on a direct-response basis, and then wonder why branded search volume drops.

The fourth is short time horizons. Marketing effects, particularly brand effects, accumulate over time. A quarterly analysis cycle that only looks at in-period performance will systematically undervalue activities with longer payback periods. This is one of the reasons brand investment gets cut in downturns: the analysis framework is not built to capture its value.

Understanding why go-to-market execution feels harder than it used to is partly a measurement problem. When the analysis framework does not capture the full picture, teams make decisions based on incomplete information and then wonder why growth is harder to sustain than the dashboards suggested it should be.

How to Build the Framework Without Starting From Scratch

Most teams already have the raw material for a better analysis framework. The data exists. The problem is usually in how it is organised, interpreted, and acted upon.

Start by auditing what you are currently measuring and asking whether each metric is connected to a commercial outcome. If you cannot draw a clear line from a metric to revenue, margin, or customer growth, it is either a diagnostic metric (useful for understanding why something is happening) or a vanity metric (useful for making slides look good). Know which is which.

Then identify the gaps. Which of the four layers are you missing? Most teams will find they have reasonable channel performance data, weak audience composition data, almost no market context layer, and a commercial outcome layer that is disconnected from the marketing analysis entirely.

Close the gaps incrementally. You do not need a perfect framework on day one. You need a framework that is honest about its limitations and improves over time. An imperfect framework that surfaces the right questions is more valuable than a polished one that only confirms existing assumptions.

The broader thinking on go-to-market and growth strategy that sits behind this kind of framework is something I write about regularly. If this article is useful, the growth strategy section of The Marketing Juice covers the commercial and strategic context that makes analysis frameworks actually actionable.

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