KPI Metrics: Stop Measuring What’s Easy, Start Measuring What Matters

KPI metrics are the measurements a business uses to track progress toward its most important commercial objectives. The word “key” is doing a lot of heavy lifting in that definition, because most organisations are not measuring key performance indicators at all. They are measuring convenient ones.

The difference between a metric and a KPI is not technical. It is strategic. A metric tells you what happened. A KPI tells you whether the business is moving in the right direction. Most dashboards are full of the former dressed up as the latter.

Why Most KPI Frameworks Fail Before They Start

I have sat in more measurement planning sessions than I can count, across agencies, client-side teams, and board-level strategy reviews. The pattern is almost always the same. Someone opens a spreadsheet, lists every metric the business can currently track, and then labels the most important-sounding ones as KPIs. That is not a KPI framework. That is a data inventory with ambitions.

The failure happens at the starting point. Most organisations begin with the question “what can we measure?” when they should begin with “what are we trying to achieve, and what would tell us we are getting there?” Those are fundamentally different questions, and they produce fundamentally different outputs.

When I was running iProspect and growing the team from around 20 people to over 100, one of the most consistent challenges was getting clients to separate their reporting layer from their decision-making layer. They wanted dashboards that showed everything. What they needed was a framework that told them three or four things with clarity. The more metrics on the screen, the more comfortable people felt, and the less clearly anyone could articulate what was actually working.

If you are building or rebuilding a measurement framework, the Marketing Analytics hub on The Marketing Juice covers the broader infrastructure behind this, from GA4 configuration to attribution modelling. The KPI question sits on top of that infrastructure, and it is worth getting the strategic layer right before you configure the technical one.

What Makes a Metric a KPI

There are four criteria a metric needs to meet before it earns the label of KPI. Most metrics fail at least one of them.

First, it must be tied to a specific business objective. Not a marketing objective. Not a channel objective. A business objective. Revenue, margin, customer acquisition, retention, market share. If you cannot draw a direct line from the metric to one of those outcomes, it is not a KPI.

Second, it must be actionable. If the number moves and no one knows what to do differently, it is not a KPI. It is an observation. Actionability means the metric creates a decision trigger. If cost per acquisition rises above a threshold, you do something specific. If conversion rate drops, you investigate a specific set of causes. The metric has to connect to behaviour.

Third, it must be owned. KPIs without owners are decorative. Someone in the organisation needs to be accountable for each indicator, with the authority to influence it. This is where a lot of marketing KPI frameworks break down. Organic traffic might be a meaningful metric, but if no one on the team has a mandate to do anything about it, it is not a KPI worth tracking at the executive level.

Fourth, it must be time-bound. A KPI needs a target and a timeframe. “Improve conversion rate” is not a KPI. “Increase landing page conversion rate from 2.1% to 3.0% by end of Q3” is a KPI. The specificity is not pedantry. It is what makes the metric useful for decision-making rather than retrospective storytelling.

The Vanity Metric Problem Is Worse Than You Think

Vanity metrics are not just a problem for junior marketers who are excited about Instagram follower counts. They exist at every level of an organisation, often dressed in more sophisticated clothing.

Impressions, reach, share of voice, brand search volume, email open rates, webinar registrations, social engagement rate. None of these are inherently useless. All of them become vanity metrics the moment they are reported without context, without a commercial connection, and without a decision attached to them.

When I was judging the Effie Awards, one of the things that separated the strong entries from the weak ones was not the scale of the results. It was whether the team could articulate what the metrics meant for the business, not just for the campaign. Plenty of entries had impressive reach numbers and award-worthy creative. The ones that stood out could connect the work to commercial outcomes with evidence, not assertion.

Email marketing is a good example of where vanity metrics can quietly dominate a reporting framework. Open rates feel meaningful. Click-through rates feel meaningful. But if your email programme is not connected to revenue attribution, you are measuring the health of your email list, not the performance of your marketing. The CrazyEgg breakdown of email marketing metrics is a useful reference for thinking about which email indicators actually connect to outcomes versus which ones just confirm that people received your message.

The deeper problem with vanity metrics is that they are self-reinforcing. When a team is measured on reach, they optimise for reach. When they are measured on engagement, they produce content that generates engagement. Neither of those things is wrong in itself. The problem is when reach and engagement become proxies for business performance without anyone testing whether that proxy relationship actually holds.

How to Choose the Right KPIs for Your Marketing Function

There is no universal KPI framework that works across every business, channel, or stage of growth. Anyone selling you one is selling you a template, not a strategy. The right KPIs depend on what the business is trying to do, what stage it is at, and what levers marketing actually controls.

That said, there are useful categories to work from. Most marketing KPI frameworks need to cover three levels: business-level outcomes, marketing-level outcomes, and channel-level indicators.

Business-level outcomes are the metrics that appear in board reports. Revenue, profit, customer lifetime value, churn rate, market share. Marketing does not own all of these, but it should be able to articulate its contribution to each of them. If marketing cannot explain how its activity connects to any of these numbers, that is a strategic problem, not a measurement problem.

Marketing-level outcomes sit one step down. Cost per acquisition, return on ad spend, lead-to-customer conversion rate, customer acquisition cost by channel, share of new customers from marketing-attributed sources. These are the metrics marketing owns directly and should be held accountable for.

Channel-level indicators are the operational metrics that explain why the marketing-level outcomes are moving. Click-through rate, quality score, bounce rate, time on page, video completion rate. These are diagnostic tools, not KPIs in the strategic sense. They tell you where to look when something changes. They should not be the primary metrics in a board-level conversation.

Content marketing is an area where this three-level distinction gets blurred most often. Teams track page views, social shares, and time on site without connecting those numbers to pipeline or revenue. The Semrush content marketing metrics guide is worth reading for a more structured approach to connecting content performance to commercial outcomes, and the Unbounce breakdown of content marketing metrics covers the diagnostic layer in useful detail.

The Attribution Problem Every KPI Framework Has to Solve

Attribution is the point where most KPI frameworks quietly fall apart. The question of which channel, touchpoint, or campaign gets credit for a conversion is not a technical question. It is a political one, and it shapes which metrics get elevated to KPI status.

Last-click attribution, which remains the default in many organisations despite being widely criticised, systematically overstates the contribution of bottom-of-funnel channels and understates the contribution of everything that builds awareness and consideration. This distorts KPI frameworks in predictable ways. Paid search looks like a hero. Brand and content look like costs. Budgets follow the metrics, and the metrics are wrong.

I spent years managing paid search at scale, across hundreds of millions in spend, and the attribution conversation came up in almost every client relationship. The clients who understood attribution clearly made better decisions about channel mix. The ones who trusted their platform dashboards without question consistently over-invested in retargeting and under-invested in acquisition.

Proper UTM tracking is a prerequisite for any attribution work that goes beyond platform-reported data. If you are not tagging your campaigns consistently, your attribution data is unreliable regardless of which model you use. The Semrush guide to UTM tracking in Google Analytics covers the mechanics of this well. Getting the tagging right is the unglamorous work that makes the KPI framework trustworthy.

For teams working with GA4, the platform’s data model is different enough from Universal Analytics that KPI frameworks built on old reporting assumptions need to be rebuilt, not migrated. Custom event tracking in GA4 gives you more flexibility than UA did, but it requires more intentional configuration. The Moz guide to GA4 custom event tracking is a useful starting point for teams that need to instrument their measurement layer properly before they can trust their KPI data.

Leading Indicators Versus Lagging Indicators

One of the most practical distinctions in KPI design is the difference between leading and lagging indicators. Most KPI frameworks are dominated by lagging indicators, which measure what has already happened. Revenue, conversions, customer acquisition cost. These are important, but they tell you about the past. By the time they move, the decisions that caused them to move were made weeks or months ago.

Leading indicators are metrics that predict future performance. They move before the lagging indicators do, which means they give you time to intervene. The challenge is that leading indicators are harder to identify and validate. You need to test whether the relationship between the leading indicator and the lagging outcome actually holds in your specific context.

In a SaaS business, free trial sign-ups might be a leading indicator for paid conversions. In an e-commerce business, add-to-cart rate might be a leading indicator for revenue. In a B2B services business, qualified discovery calls booked might be a leading indicator for closed deals. None of these relationships are universal. They need to be validated against your own data before you build a KPI framework around them.

When I was working with a loss-making agency early in my career as a turnaround exercise, one of the first things I did was identify the leading indicators for revenue. The lagging data told us we had a revenue problem. The leading indicators told us where in the pipeline the problem was originating. That distinction changed the entire intervention strategy. Without the leading indicators, we would have been optimising the wrong thing.

How Many KPIs Is Too Many

There is no precise answer, but there is a useful principle. If you have more KPIs than you have people who own them, you have too many. If your KPI dashboard takes longer than five minutes to review in a weekly meeting, you have too many. If adding a new KPI does not remove an old one, you are building a reporting library, not a performance framework.

The pressure to add metrics is constant in marketing organisations. New channels create new metrics. New tools surface new data. New stakeholders ask for new reports. The discipline is in saying no to most of it, and being clear about why.

A useful rule of thumb is three to five KPIs at the marketing function level, with a small number of supporting indicators beneath each one. The supporting indicators are the diagnostic layer. They explain movement in the KPIs. They do not carry the same weight in performance conversations.

For teams running webinar or event-based marketing, the temptation to track every engagement metric is particularly strong because the data is so granular. Registration rate, attendance rate, replay views, post-webinar conversion rate. All of these are interesting. Very few of them are KPIs. The Wistia guide to webinar marketing metrics does a reasonable job of separating the diagnostic layer from the outcome layer for this specific channel.

Benchmarks, Targets, and the Danger of Industry Averages

KPIs need targets to be meaningful. A metric without a target is just a number. But where those targets come from matters enormously, and industry benchmarks are a much weaker foundation than most teams assume.

Industry averages aggregate across businesses with different models, different audiences, different price points, and different levels of marketing maturity. A conversion rate benchmark for e-commerce covers everything from luxury goods to commodity products, from first-time buyers to repeat customers, from mobile-first experiences to desktop-optimised ones. The average tells you almost nothing useful about what your conversion rate should be.

The most useful benchmarks are internal ones. What was your conversion rate last quarter? Last year? What changed, and why? Your own historical data is a more reliable baseline than an industry average, because it controls for all the variables that make your business different from everyone else in the sector.

External benchmarks are useful for one specific purpose: identifying whether you are operating in a fundamentally different range from the market, which might indicate a structural problem or a structural advantage. If your cost per acquisition is three times the industry average and your product pricing is comparable to competitors, that is worth investigating. But if you are within a reasonable range, optimising toward the benchmark rather than toward your own historical improvement is often the wrong move.

Content marketing metrics are particularly susceptible to benchmark obsession. Teams chase average session duration, average pages per session, and average bounce rate figures that have been recycled across the industry for years without much interrogation of whether they actually correlate with commercial outcomes. The Buffer overview of content marketing metrics takes a more grounded approach to this question, and it is worth reading alongside the Semrush content metrics framework for a fuller picture.

When KPIs Need to Change

KPI frameworks are not permanent. The metrics that matter at the start of a business are not the same ones that matter at scale. The metrics that matter in a growth phase are not the same ones that matter in a profitability phase. One of the most common mistakes I see in mature marketing organisations is carrying forward KPIs that made sense three years ago but no longer reflect the current strategic priority.

A business in early growth might rightly prioritise customer acquisition above all else, accepting a high cost per acquisition in exchange for market share. As the business matures, the KPI framework should shift toward retention, lifetime value, and margin. If the measurement framework does not evolve with the business strategy, the team will keep optimising for the wrong things long after the strategic context has changed.

The trigger for reviewing a KPI framework is not a fixed calendar schedule. It is a change in business strategy, a change in competitive context, or a recognition that the current metrics are no longer driving useful decisions. If your weekly KPI review is generating the same conversation every week with no new insight, the framework needs to change.

For teams that have moved to GA4 and are building their measurement infrastructure from the ground up, there is an opportunity to be more intentional about KPI design than was possible in Universal Analytics. The ability to export raw event data to BigQuery, as covered in the Moz Whiteboard Friday on GA4 and BigQuery, opens up more sophisticated analysis of leading indicators and attribution that simply was not accessible to most teams before.

Building a KPI Framework That Actually Gets Used

The best KPI framework is the one that gets reviewed weekly, informs real decisions, and changes behaviour. A beautifully designed dashboard that no one acts on is a reporting exercise, not a performance management tool.

Getting there requires a few things that are less about measurement and more about organisational design. First, the KPIs have to be visible to the people who can influence them. If the team running paid search cannot see their cost per acquisition data in real time, the KPI is not functioning as a decision tool. Second, the review cadence has to match the pace at which the metrics move. Weekly for channel-level indicators, monthly for marketing-level outcomes, quarterly for business-level outcomes is a reasonable starting structure for most organisations.

Third, and most importantly, the conversation around KPIs has to be about decisions, not just numbers. “Our conversion rate dropped 0.3% this week” is a data point. “Our conversion rate dropped 0.3% this week, we think it is related to the mobile checkout change we deployed on Tuesday, and we are reverting it today” is a KPI conversation. The metric only earns its place in the framework when it is consistently generating that second kind of conversation.

Early in my career, before I had a budget for anything, I learned to build things myself rather than wait for resources to appear. That instinct carried through into how I think about measurement. You do not need a sophisticated analytics stack to have a functional KPI framework. You need clarity about what you are trying to achieve, honesty about what the data is actually telling you, and the discipline to act on it. The tools matter less than the thinking.

If you want to go deeper on the analytics infrastructure that sits beneath a KPI framework, the Marketing Analytics section of The Marketing Juice covers GA4 configuration, attribution, and measurement strategy in more detail. The KPI layer is only as good as the data layer beneath 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