Blog / How to Change Board Reporting Metrics Without Moving Goalposts

I’ve worked with so many marketing leaders over the years whose board has been looking at the same number for as long as anyone can remember because it’s familiar. Form fills. MQLs. Organic traffic. Whatever it happened to be when the dashboard was built.

But in the last 18 months our industry has experienced some of its biggest shifts and nearly all of them have reached the same uncomfortable conclusion. The number they used as a target has stopped meaning what it used to mean, and the moment they say so out loud they are going to look like someone trying to wriggle out of a target.

One of these marketing leaders said it to us earlier this year. We’d proposed adding an AI visibility measure alongside the organic traffic goal an engineering software company had been reporting on for years. The person who owns that reporting internally pushed back. “You can kind of move the goal post however you want at any time if we don’t have a standard way of what is in scope.”

He was telling us, quite reasonably, what it was going to look like to the execs he had to present this change to, and he was right.

This is what we’ve learned about making that change and surviving it. I think the procedure matters a lot more than which metric you pick.

Why Marketing Metrics and Organic Search Benchmarks Shifted

The search lead at the same client stated the problem better than I could. “Is there less demand for the brand, or is AI picking up?” Their organic traffic was going down, but the company was showing up more than ever in AI answers, and the reporting for those two things lived in different places, so nobody could say which story was true.

The board was only being shown the traffic line, which was the half that looked like decline. It’s a horrible position to be in. I’ve lost count of how many times a client has said “our organic traffic is down” to us in the last three months. It’s ingrained, this looking at traffic, because traffic always went up, so we looked at it, and the board learned to look at it too.

It goes deeper than AI, though. When we’ve looked under the hood at a traffic number, a lot of it was never going to convert. It could be people coming to log in. It could be people looking for the address. There could be a whole host of reasons, and all of it is diluting your real conversion rate, which is the rate among the people who arrived with some commercial intent. So when a top of funnel paid campaign goes live, or an informational blog post starts ranking, it looks like the conversion rate dropped, when the reality is all that extra traffic was never going to convert anyway. It’s only when you’ve got those segments that you can weather the storm and actually explain it through the business.

Then there’s the tracking itself. The data is typically going from GA or Adobe into Salesforce, or similar. Consent mode adds another layer, where a good share of your users don’t exist as far as the analytics is concerned. There is so much opportunity for it to become unreliable that people simply don’t trust it, or don’t have the budget or the skill set to fix it. At an identity software company we found 576 conversions in the ad platform for a month where the CRM had 148. Same month, same campaigns, and for years the platform number had been THE number. Nobody had checked because it was “looking good”.

Why Leadership and Board Reporting Demand Transparency

To be fair to the board, the suspicion is earned. Gartner surveyed 125 CEOs and CFOs in late 2024, and only 22% of them said their CMO had given them real clarity on what marketing was accountable for. If you don’t know what a number means, you can’t tell a genuine redefinition from a rescue, so you assume rescue. I don’t blame them. Nobody explained the scoreboard to them in the first place, and I think a lot of the reason marketing isn’t taken seriously is that it has spent years reporting marketing metrics the board doesn’t actually care about.

The marketing leaders I deal with know this, by the way. They know the number is often unreliable. But they also know what they’re going to be measured on, and if it moves there are going to be a lot of questions asked. This is why they become anxious, knowing leadership will ask these targeted questions.

The OKR crowd have a name for the bad version. They say the bar for changing a measure mid-cycle should be high, and that changing it because you’re afraid of missing is sandbagging.

No argument there. We watched one client go from optimizing to return on ad spend, to offline conversions (volume fell by roughly 90%), to qualified leads, and back again over a couple of years. Their own words were that they were continuously jumping from one to the other. Every change was a swap, nobody wrote down where they started, and each change happened because the last one had produced a problem.

No number survived long enough to mean anything and this is why boards lose confidence.

So the company-level education piece of what is important to truly measure success still has to happen first, which is kind of unbelievable in 2026, but it’s true. Form fills mean nothing if they’re not becoming MQLs. The MQLs mean nothing if they don’t become SQLs, and SQLs don’t mean anything if they don’t become pipelines. Once everyone accepts that, you can change the number, and if you do it properly the person who said “you could move the goalposts however you want” ends up defending the new one for you.

Five Strategies for Changing KPIs without Losing Trust

Here are five moves that help you have conversations about changing metrics from what is familiar to what is useful with the wider company:

1. Keep the old number on the slide and put the new one beside it. The line our operations director used to close the argument above was “we’re not moving the goalposts, we’re just adding more metrics.” The organic traffic number stayed. The AI visibility number went next to it. That costs you something, because you keep being graded on the old line for a while, and the cost is why anyone believes it. If you swap the number, it gets read as an escape, but if you put a second number next to the first, you’ve taken on more accountability, and people reading it can see that, which I think is why that particular argument closed as quickly as it did.
   
   
2. Fix the baseline and the source before anyone sees the new number. On a website consolidation this summer we ran three sets of data in parallel. The old domain, the new one, and a third property we deliberately left alone so we had something that shouldn’t move. In week one the migrated estate showed a small decline and the property we hadn’t touched had moved MORE. One sentence, and nobody was frightened of the number any more. On another account the client’s dashboard showed 3 form submissions for a product where their analytics tool showed 11, and both were right, because they were counting different things with different filters. Agree what the number is and where it comes from first, or you’ll spend the meeting arguing about whose number is right instead of what to do about it
   
   
3. Let the other party own the instrument. This is the one most agencies won’t do. On one account we were reporting 22% AI citation share and the client’s own tool said 34%, for the same thing, because we were tracking different prompts in different tools. So we gave up ours and worked inside theirs. (The tool, it turned out, had shipped with no competitors configured, so it had been ranking a B2B software company against Wikipedia and Reddit. Nobody at the vendor had thought to mention this). On the engineering software account we conceded it up front, “you guys can own the tools, it’s fine.” If you’re the marketing leader, your version of this is to let finance or revenue operations hold the instrument the board number comes from. You lose the ability to pick the tool, which is a small price for being believed
   
   
4. Write the definition down for the person who joins next year. After the 3 versus 11 conversation, the client asked for the logic behind the report in a format anybody could pick up, because there was no context around how it had been built. What counts as a form submission, which pages are excluded, which prompts are tracked, which competitors, which platforms. We now sign that off row by row on a spreadsheet before anything is reported, and the definition sits beside the number on the slide. It’s boring. I’ve also watched it be the only reason a number survived a meeting I wasn’t in.
   
   
5.  Say what is going to fall before it falls. We told the identity software company, before a signup change went live, that reported conversions would drop by roughly 200 a month and that nothing would drop in the CRM. It did exactly that, and the client asked for that note to live permanently on the report so nobody arriving late read it as a decline. That’s a forecast you can be graded on.

Applying Financial Principles to Marketing Accountability

I’ll admit this one annoyed me slightly. Finance has had the answer written down for a decade, and accounting draws a line marketing never draws, between a change of policy and a change of estimate. A change of estimate means the world you were measuring has moved and you’ve updated the number to match, so you apply it going forwards, you keep the old series, and you show the two together. That is exactly what has happened to organic traffic in 2026. Organic traffic was a fine number for the world it was built in. That world has moved, so you show the new number and the comparison.

The European guidance on alternative performance measures, which is any number a company chooses to report that isn’t in the statutory accounts (EBITDA, like-for-like sales, that sort of thing), was issued in October 2015 and has applied to anything published since July 2016. If you redefine one, it tells you to explain the change, explain why the new version gives people more reliable and more relevant information, and provide restated comparative figures. The SEC runs along the same lines, asking you to disclose the change, explain the reasons for it, and recast the earlier periods where the change is significant enough to matter. It’s rare to find anyone in marketing who has read either, and it is our procedure almost word for word.

Building AI Visibility and Measurement Trust

Our Head of SEO did more for trust on the engineering software account than any dashboard, by saying, unprompted, “I don’t know if this is allowed, but I also don’t trust any of the AI tracking tools.” And the client’s response was “no, but that’s good, because we can present it that way.”

Admitting the new instrument was imperfect gave him something he could say upward. I think the board wants someone to tell them the rules before the game starts, and then to play by them, and they’ll forgive an imperfect number a lot faster than a surprise.

Aligning Marketing Metrics with Finance for Board Reporting

There still has to be a number a board can hold you to, and mine is a coarse one. At the end of each quarter and each year, this is what we spent on media, this is what we spent on marketing including agencies and staff, and this is what the business generated, and you should be seeing a return on it. Everything in the procedure above exists so you can walk into that room and say, hey, we’ve got fifty percent less form fills, but we’re able to show fifty percent more revenue, and be believed.

Agree that framework with finance before the budget is allocated. Every marketing leader I’ve watched hold a CFO’s confidence that agreed early, while the old number was still going up.