A 30-minute check on whether you can trust your marketing numbers

Before you change a campaign, it’s worth knowing whether the numbers you’re about to act on are real. This check takes about half an hour and needs no special tools: a browser, your ad accounts, your analytics and your CRM.

You are looking for one thing. Do these systems tell the same story about the same week?

1. Pick one week and count leads three ways

Choose a full week that ended at least ten days ago, so late conversions have settled. Write down three numbers for that week:

• Conversions reported by your ad platforms, added together.
• Conversions reported by your analytics tool.
• New leads created in your CRM.

They will never match exactly. Ad platforms each claim credit for the same lead, and some people block tracking. What matters is the size of the gap. If the ad platforms claim twice what the CRM received, the platforms are being optimised towards something that isn’t a lead.

2. Submit your own form and follow it

Open your site in a private window, click through from a real ad if you can, and fill in your main form with a name you’ll recognise. Then go looking for yourself.

Did the conversion appear in analytics? Did it appear in the ad platform the click came from? Did the lead arrive in the CRM, and does the record say where it came from?

If the source field is empty, or says “direct” when you know you clicked an ad, the link between marketing and sales is broken at that point. This one test will show you more than any dashboard.

3. Check what counts as a conversion

In each ad platform, open the list of conversion actions and read it line by line. You are looking for three common faults:

• Page views or button clicks counted as conversions alongside real form submissions.
• The same form counted twice, once by an old tag and once by a new one.
• Conversions that fire when the form loads, before anyone submits it.

Any of these will make a campaign look better than it is, and the platform will spend your money chasing the cheap fake signal.

4. Look at your links

Open five recent ads or emails and look at the destination links. Are they tagged so analytics can tell campaigns apart? Are the tags consistent, or does one campaign appear under four slightly different names?

Inconsistent naming doesn’t lose data. It scatters it, so no report ever adds up. A one-page naming rule fixes it from that day forward.

5. Ask where revenue lives

Finally, ask a plain question. When a lead becomes a customer, where is that recorded, and can it be traced back to the campaign?

If the answer is “in the finance system, and no”, you can still run good marketing. You’ll be steering by cost per lead, though, and the cheapest leads can be the worst ones. Knowing that limit is useful in itself.

What to fix first

Fix in the order the data flows. A broken form-to-CRM link comes first, because every later number depends on it. Conversion definitions come second, because they steer the ad platforms’ spending today. Link naming comes third. Revenue tracking is the largest job and the most valuable, and it goes much faster once the first three are clean.

None of this is glamorous work. It is the difference between changing a campaign because of what happened and changing it because of what a miscounting tag said happened.

If your three numbers in step one were far apart and you’d like a second pair of eyes, this is what our two-week tracking sprint covers.