Why nobody trusts your dashboard.

Why does nobody use the dashboard we paid for?

Because it can't be trusted, and an untrusted dashboard is just an expensive screensaver. You've seen it happen. Two people pull "the numbers" for the same month, land on two different totals, and the meeting quietly stops being about decisions and starts being about whose spreadsheet is right. Nothing gets decided. Everyone goes back to gut feel.

Here's the thing though: that's almost never a charting problem. Nobody's dashboard is broken because the bar chart is the wrong shade of blue. It's broken because the numbers underneath don't mean what people think they mean. It's a trust problem and a definitions problem. Fix those and the charts sort themselves out.

The symptom: two reports, two numbers, zero decisions

The tell is always the same. Marketing says the campaign drove 50 conversions. Sales opens the CRM and sees maybe 20 real people. The CEO watches the two of you disagree and reasonably concludes the whole thing is noise, so they stop looking at it. Once a leader stops trusting a dashboard, they never fully come back to it, and you're running the business on vibes again.

So let's talk about what's actually rotting underneath. In my experience, a dashboard lies about two things more than anything else: conversions and traffic. Get those two honest and you've closed most of the trust gap.

Why is your dashboard lying to you?

Because it's counting things that aren't real, and mixing in people who were never going to buy. Two culprits, both fixable.

The conversions problem: double-counting and spam

Two ways conversions get inflated, and I see both constantly.

Double-counting. Usually it's two pixels firing where there should be one, or you're tracking both a button click and a success-page load for the same form. One person fills in one form, and your analytics happily records it twice. Suddenly your conversions have doubled and everyone's thrilled, "wow, we're crushing it", when actually nothing changed except the maths. It's the easiest way in the world to accidentally lie to your own CEO.

Counting spam. This is the sneakier one. You proudly report 50 form fills. But depending on who you're presenting to, say, a CEO of a B2B business, the real question isn't "how many things touched the form?" It's "how many actual humans did we convert?" Of those 50, maybe 20 are valid. The other 30 are bots, people after a job, or someone pitching you their agency services. They're not conversions. Reporting all 50 feels great and tells nobody the truth.

"A CEO doesn't care that fifty things filled out your contact form. Half of them were bots, job-seekers, or someone trying to sell you their services. They want to know how many real humans you actually converted."

So don't report the 50. Report the 20 a human would recognise as genuine. Fewer, honest numbers beat big, hollow ones every single time.

The traffic problem: no filtering and the wrong region

Traffic gets inflated the same way, by counting people who were never in the running. The classic version: you only operate in the UK, but you present global traffic to the whole company. Now your headline number includes visitors from Singapore, Australia, Canada and Argentina, lovely people, but not one of them can buy your product. (Sure, a handful are UK folks on a VPN who forgot to switch it off. That's the rounding error, not the point.)

If you sell in the UK and you're chasing UK buyers, build the dashboard around UK traffic. Make it focused and intent-led, and ignore the rest. Keep a secondary view somewhere to catch genuine surprises. "Hey, we've suddenly got a spike from over there, maybe that's a market worth going into." But your main dashboard, the one you use to check on your actual audience and your actual progress, should only show the people who could ever say yes.

How do you rebuild trust in a dashboard?

You rebuild it in a specific order, and skipping steps is why most "dashboard refreshes" fail. It goes: instrument, define, validate, then one source of truth.

• Instrument cleanly. Fix the tracking first. One conversion event per action, no duplicate pixels, no button-click-plus-success-page double count. Get the raw signal honest before you build anything pretty on top of it.
• Define every metric. Agree, in plain words, what each number means. What counts as a conversion? What's a lead versus an MQL? Write it down where sales, marketing and the founder can all see it. Most "two reports, two numbers" fights are just two people using the same word for different things.
• Validate against reality. Cross-check the dashboard against the CRM and against common sense. If the dashboard says 50 and the CRM says 20 qualified, that gap is the story, go find it before you present it.
• Run one source of truth. Pick the system everyone reads from and make it the canonical one. Tools like GA4, Looker Studio (the old Data Studio) or HubSpot are all fine, what matters is that there's one agreed home for the number, not five.

None of that is rocket science. It's mostly the discipline to define things before you chart them, which is exactly the sort of analytics and tracking groundwork we do before we'll let a dashboard anywhere near a boardroom.

Which metrics should you actually trust?

The ones tied to real intent and real humans, and, above all, the ones you watch as a trend rather than a single sacred number. Here's my personal shortlist.

• MQLs under a strict definition, not raw leads, not every contact-form fill. A marketing qualified lead is a real human whose message was legitimate and who fits the criteria we agreed. That's the number I'll take to a founder.
• "Read more" expansion clicks on LinkedIn, someone actively choosing to expand and read our post is real brand engagement. Impressions aren't; you can buy as many of those as you like, any day of the week, and most of them are people scrolling straight past.
• Traffic from my actual target region, UK traffic if the UK's the market. Not all-traffic-everywhere.
• Trends over absolute numbers, this is the big one.

"I'll take a trend I trust over a number I can't every time. If ChatGPT referrals go from 14 to 50, and overall UK traffic keeps climbing, I don't need the tracking to be perfect, I can see AEO is working. It's the trajectory I care about, not the verbatim figure."

Referral tracking for AI and social is famously leaky. People see you on ChatGPT, then Google you directly, and the credit lands in the wrong bucket. So I don't treat those raw numbers as gospel. I watch the line. If the overall trend is up and to the right, something good is happening, even if no single tool can prove exactly what. Raw impressions and all-traffic-everywhere I treat as vanity. Trend lines tied to real intent, I trust.

Get your conversions honest, filter your traffic to people who can actually buy, define your terms, and watch the trend, and suddenly people start using the dashboard again. If you'd like senior operators to rebuild yours so the whole company trusts the number on the screen, that's very much our thing. Have a look at what we do with analytics and data and the tools behind it, or just grab a quote.