Article

Reporting Needs A Revolution

Advertising Analytics Data Marketing reporting

by Sam Tomlinson

September 15, 2025

This week, I want to talk about an area that receives precious little attention – but demands massive amounts of time from most marketing teams/agencies: reporting. I was genuinely gobsmacked when I spoke to two fellow mid/large agency owners, who both told me their teams spend 8-10 hours, per month, per retainer client on reporting (data pulls, report assembly, actual meeting with client to review the reporting, post-meeting follow-ups). I thought – for sure! – this was an anomaly….so I talked to another 10 people (in house + agency) about it, and they all came back with similar figures when I showed them the itemized breakdowns.

If you run an agency with 10 employees and 15 retainer clients, that implies you might be spending 150+ staff hours per month just on reporting! That’s an FTE’s entire capacity and then some.

Before we go further, I want to be clear: just because something takes a decent chunk of time does not mean it is inherently a problem. Those teams likely spend similar (or even greater) portions of their time working in ad accounts, making new ads, optimizing post-click experiences, etc. It’s only a problem if the return from the investment (time/resources) does not justify the cost.

And THAT is the heart of the problem. When I look at reporting, there’s a power law at work: >80% of the actions/improvements are driven by <20% of the reporting. The logical implication of that is that 80% of the reporting – 6.4 to 8.0 hours per client, per month – is spent on the reporting that drives <20% of the impact.

That’s categorically insane. And for marketing to be better, it needs to change.

Every hour spent on one task is an hour that is unavailable to be used on another task. An hour spent on reporting is an hour that can’t be used on account updates or new creatives. An hour spent on building a dashboard is an hour that can’t be used building a new automation. Time is like money: it’s zero-sum.

That makes protecting + optimizing your time one of the highest leverage things anyone – in-house, agency, freelancer, whatever – can do.

As I’ve spent more time thinking about reporting – what it is, what it ought to be – I’ve come to three conclusions that shape my approach to it:

1. Reporting is only valuable if it contributes to future positive outcomes – whether that’s understanding a mistake (thereby enabling you to avoid it in the future) or informing a smarter future course of action
   ​
2. The scope of reporting should be limited to the data points that enable (1). Too many of the reports/dashboards I see are 80% data pukes / 20% useful insights + analysis.
   ​
3. Attention is light: focused into a tight beam, it can cut through the hardest materials in the known universe; diffused, it will barely increase the ambient temperature a degree. Everything we do from a reporting perspective should focus attention on the most critical things, not diffuse it across a thousand charts/graphics/metrics.

When I share this with other CMOs, (2) is the point that receives the most pushback – most leaders love to see all the data. They fancy themselves a modern-day Marketing Indiana Jones, who is going to dig around and find the golden nugget of insight that all of the supposedly-smart people they’ve hired missed (though I can count on one hand the number of times a CMO/VP has actually done that and surfaced a useful insight, in my 15+ years). The reality is that this just doesn’t happen – most leaders have too many demands on their time as it is, and even if they did, they don’t actually have the expertise, contextual knowledge or supporting skills (i.e., coding or data science knowledge) to do it.

Instead, what ends up happening is overly long, hyper-detailed reports are built that serve no purpose but to collect dust and take up harddrive/cloud storage space. Reporting meetings devolve into data puking sessions that could put a ferret on cocaine into REM sleep. Nothing changes. Everyone eventually gets surprised that nothing has changed. Sadness (or firings) ensue.

The conclusion I’ve come to is this: reporting is a gas. It expands to fill the space it’s permitted to take. And too often, we’ve allowed that gas to expand to progressively larger spaces, without asking several critical questions: Why are we reporting on this? How does this information help us going forward? What should we do next as a result?

That’s all that’s relevant.

What we need is a revolution in how the entire marketing ecosystem approaches reporting. This starts with two paradigm-shifting changes: (1) anchor everything from a reporting perspective around four questions:

1. Why did this happen?
2. How should we respond/react going forward?
3. What impact will doing that have on our business/goals?
4. When will that impact materialize and in what form?

And (2) commit to flipping the distribution of your reporting activities, such that 80%+ of your reporting is leading to action.

Doing just those two things can (and will) have a transformative impact on your business (to say nothing of your agency’s sanity + your team’s happiness).

If you’re with me so far, you’re probably wondering how to accomplish this. What are the concrete steps that take you from your current reporting to this new paradigm?

You’re in luck. Here’s exactly how I’m thinking about doing that:

1. Identify your core KPIs
2. Automate + Dashboard
3. Trend + Segment
4. Forecast + Benchmark
5. Connect + Prioritize
6. Execute
7. Evaluate

Identify Your Core KPIs.

Marketers have been too obsessed with vanity for too long. That’s led to stuffing reports + dashboards with feel-good metrics that go up-and-to-the-right, but don’t provide insight into the health of the underlying business.

That must end for reporting to improve.

Start by identifying the KPIs that actually matter to your (or your client’s) business. For reference: a KPI is a metric that maps to a critical business function and enables understanding of the business’ performance vs. your objectives/goals/targets.

I’ve used the below slide to communicate this idea before:

Once you’ve identified your KPIs, delete everything else.

Why?

Because everything that you’ve deleted is a metric. It’s informative for lever-pulling and small-scale optimization, but it’s not something that will help you decide what to do going forward. The people on your team or your agency’s team can use metrics to improve their day-to-day activities, but reporting on them will not lead to better outcomes or results – to the contrary, it will actually siphon attention away from the things that really matter. Reporting on bullshit is not an innocent or harmless activity; to the contrary, it actively undermines the true goal of reporting.

Focus your reporting on what actually matters to the business. Deep dives, breakdowns, etc. can be done on an ad hoc basis, when a specific situation warrants it.

Automate + Dashboard

Time is the most valuable asset any expert has – but far too many waste far too much of it on manual tasks like data pulls.

Stop it.

We build every client a dashboard (or series of dashboards, depending on the brand) at the outset of our relationship. And, to be candid, we do it because it’s in our best interest: I don’t want our team expending precious time to pull a rudimentary report. I don’t want our clients waiting around for said report, or feeling like they need to ask for it. Dashboards solve both problems – our team doesn’t have to spend time pulling data (it’s done automatically), and our clients can ALWAYS see what’s going on in their accounts. Transparency, accountability + efficiency.

Yes, there are absolutely some clients who have never logged into their dashboard. It happens. And in those cases, we adapt and evolve to ensure they are getting what they need. But, in my experience, allowing the exceptions to drive the SOPs is a terrible idea.

The reality is that using dashboards has saved our team thousands of hours over the last few years (very likely more!) AND the majority of clients LOVE their dashboards. They log into them at all hours of the day. They ask questions based on what they’re seeing in their dashboard! And, on the flip side, our team is more productive than ever, because they’re not having to pull hundreds of ad hoc reports every year.

Much of the pushback I get to this is optics-related: “won’t the client think we’re not doing as much?” “clients like when reporting is personal – it shows effort!” “reporting is how we show our value to clients/stakeholders”

That’s all hogwash.

You know what shows value? Creating value.

You know what shows clients you’re doing stuff? Doing stuff that makes their business better.

Time is zero-sum. You can choose to spend time assembling pretty decks and mile-long reports, or you can spend time working on your client’s account/business. Choose wisely.

Trend + Segment

As we create dashboards, there are two absolutely essential things we do: trend + segment every KPI on which we’re reporting.

It’s human nature to over-react to the present and to discount the past. Most of the time, we forget that the same thing happened last quarter or last year. Just this week, I was talking to a client who lamented, “PPC just doesn’t work like it used to…” In this case, he was right – but not in the way he thought. His performance was actually double what it used to be, with half the volatility. The difference is perception: he’s gotten so accustomed to wins that when he goes a few days without one, it feels apocalyptic.

This is where trending KPIs is essential.

By plotting the KPI over time (days for up to 90 days; weekly for 13 weeks; monthly for 13+ months), you’ll make it easy to spot when something is within a normal range or an anomaly. This is the much-needed context and normalization that allows reporting conversations to be infinitely more productive.

As a general rule, if a KPI is within the normal variation, it’s not that interesting (unless there’s something else going on – more on that later!); if it’s anomalous, it’s worth additional investigation.

The second most common issue I see in reporting is a failure to understand Simpson’s Paradox (if you’re unaware – Simpson’s Paradox is a statistical phenomenon where the trend of data appears different when groups are combined vs. analyzed separately).

We are in the midst of an audit for a brand now, and they shared their agency’s reporting – top-level traffic, sales & revenue. Month-over-month, it seemed pretty normal – but when we dug into the underlying distribution of that traffic + sales, we saw a radically different picture: organic search was driving more and more of the sales, while paid was spending more with lower performance. The net was the same (higher organic sales offset lower paid sales), but the insight was in the underlying distribution.

The solution? Segment your data! If one KPI is CAC, segment it down by channel/campaign. If a KPI is churn, segment it by cancellation reason or cancelled service. If a KPI is appointments booked, segment by location, campaign or service type.

If you’re wondering how to segment your data, the easiest way to figure it out is to ask: if I told the client/boss this top-level KPI, what would s/he ask next?

The combination of trends (which show temporal variation) and segmentation (which reveal the movements of underlying cohorts) makes reporting 10x more actionable – it reveals if what’s happening is actually noteworthy AND what group(s) are disproportionately impacted. From there, you can move to action!

Forecast + Counterbalance

I have a rule: never show a KPI alone. This is Goodhart’s Law in action – when a metric becomes a target, it ceases to be a good metric.

Instead, show your KPIs vs. your forecast AND a reasonable benchmark (historical, industry, competitor, etc.). This allows everyone who looks at the dashboard to contextualize the KPI. We refer to these as “metric trios”. We then counterbalance each KPI with a corresponding one (also presented with forecast + benchmark).

The end result? A much clearer understanding of what’s actually going on.

A case in point: one of our clients has a KPI of cost per SQL. The value of this KPI was $512 last month.

Is that good? Bad? Mediocre? IDK.

But, if I present this data using the methodology above:

Cost / SQL: $512

Forecast Range: $551 – $674

Same Month Last Year: $793

SQLs: 82

Forecast: 63-77

Same Month Last Year: 59

Now, things are crystal clear. You know that you’ve exceeded both last year’s performance AND your 2025 forecast.

The other benefit? You can’t over-optimize one KPI.

If you just presented the cost per SQL, you could simply set super-conservative targets, spend less and make the metric look awesome. But….that would come at the expense of volume, which hurts the business overall (since you can’t take efficiency to the bank). Conversely, if you just presented total SQLs, you could simply spend more and drive the number through the roof – even if doing so resulted in SQLs coming in at costs way above what’s profitable for the business.

Presenting both together, alongside both a forecast + a historical anchor (where available), allows everyone to see what’s going on AND the natural tension between metrics (i.e. efficiency degrades with scale).

Connect + Prioritize

The thing – to me – that makes reporting magical is the ability to connect what happened to why it happened to how we move forward. Put another way, reporting is instrumentally valuable: it has value when it helps us improve going forward.

This is where marketers must be better. We must intentionally connect what happened + why it happened (what you reveal via your segments, trends + KPIs) to how to move forward.

The easiest way I’ve found to do this is ask, “So, now what?”

Take the following example, from a real client report: we’ve found that CPA has increased 19% among our adult children segment, which is 74% higher than forecasted and a ~43% increase vs. same month last year.

Ok, so what comes next? What are you going to do about it?

This forces you to confront two things: (1) why it happened (the drivers of the outcome) and (2) the path forward.

In this case, the drivers were a combination of anomalously low conversion rates and higher-than-expected costs (likely a result of the above). The recommended action was to revise the post-click experience to include the form at the top of the page AND a phone number, plus add two more CTAs at key drop of points on the page.

Expected result?

Conversion rates should revert to the previous level of 4.25% – 5.35% within the next week.

Observation → Cause → Action → Outcome

These are the kinds of conversations I want every report to inspire…because if they don’t, the report isn’t worth doing.

If this process is done well, every report should surface multiple of these opportunities – we might find under-performance among one segment, overperformance in another, an opportunity over there, etc. But time is zero-sum. Every minute we spend fixing the landing page is a minute we can’t spend on ad creative or content creation or email flows.

That’s where prioritization comes into play.

Once you’ve identified all of the relevant opportunities, go a step further: suggest what to do next, and what to do after that, and so on. You’ve now transformed something boring and low value (reporting) into something strategic and high leverage (prioritization + roadmapping).

Execute

Everything up to this point matters exactly zero IF you fail to execute. Once you’ve identified what to do, you have to do it well. It’s really that simple.

Forecasting + Reporting are exercises in operations. If you can’t execute, none of this matters.

Evaluate

The final, fatal flaw I see in most reporting is inflexibility – we report on this because we have always reported on this.

It’s absolutely insane.

Businesses change. People evolve. Priorities shift. Your reporting + KPIs absolutely must do so as well.

I have two rules to this end:

1. No more than 5 KPIs per dashboard/report
2. If it goes unused, it’s no longer useful

The first rule is often the hardest pill to swallow – but it’s essential. Reporting bloat is real. I recently went back and looked at the reports for a client over the last 6 years. The report had grown from a one-page document that focused on a core set of KPIs to a 28(!!!!) page behemoth. The time required to assemble the 28-page document was over 12 hours each month. It didn’t happen all at once – one month, the client wanted to see X. Another month, they wanted to dive into Y. Then they added another division and wanted separate reporting. Before long, one page became 6, then 6 became 11, then 11 became 14, then 14 became 19….and eventually, it ended up at 28 pages.

All because no-one was willing to cut stuff that wasn’t being used.

The real kick in the teeth? 90%+ of the stuff in that report was NEVER read.

The solution to that is the 5 KPI rule: if you want to add a KPI to a report that already has 5, then remove 1. If you can’t remove one, either (a) the thing you want to add isn’t a real KPI or (b) you don’t understand the real drivers of your business well enough, and you need to re-evaluate what you’re measuring.

And second, if you’re doing something and it’s going unused, then it’s no longer worth doing as-is. If you’re sending a client a reporting summary every week/month and you see that it is not being read/opened/used, two things need to happen: (1) have a conversation about it and (2) evolve your behavior.

Don’t be afraid of either.

There’s nothing wrong with asking a client, “Hey John, I see you opened any of the reports for the last 3 weeks. Could I send you a top-line overview of performance on Slack instead to make it easier?”

John is going to respond in one of three ways:

1. Update you on a reason why (i.e., “Hey – sorry! Jimmy is taking the lead on this now and he’s been reviewing them”)
2. Accept your suggestion (“That’d be great!”)
3. Refuse/Push back (“Hey – sorry! I don’t read these every week, but our compliance team needs them.”)

In the case of option 1, you can now reach out to Jimmy and align on what he needs for your reporting to be helpful + effective. Option 2 means you’ve likely saved an hour or so of reporting time and can help your client get the information s/he needs in the format that works. Option 3 allows you to have a more robust conversation with the compliance team and streamline/automate reporting, especially if it is being used for a different purpose (i.e., you might be able to simply automate a PDF of your dashboard to the compliance team every month, instead of spending hours pulling one together).

Then, once you get that clarity, evolve your behavior accordingly. Align how you’re sharing information to the medium/mechanism that works with your boss/client. Stop doing what didn’t work. Take that time and reinvest it into executing the things that will move your organization/business forward.

The bottom line: done well, reporting isn’t a time sink – it’s a force multiplier. The question is whether yours is multiplying value or multiplying waste.

If you follow the approach I’ve outlined, you’ll transform reporting from a time waster into a catalyst for improved performance, happier stakeholders (clients/bosses) and better outcomes.

Happy reporting!

Until next week,

Sam

Article

Optimization Outside The Ad Account

Article

How Do You Answer the “AI Strategy” Question