From the moment an ad account goes live, the platform never stops producing numbers: how many people saw it, how many clicked, what each click cost. These figures update by the hour, sit neatly ranked in a dashboard, and give the impression that something is happening. None of them, on their own, say what the business actually gained from the campaign. A click is a signal of interest, not an outcome — and when no conversion has been defined, that distinction never gets made. The gap goes unnoticed in the early weeks, because the dashboard looks full; by the time it is noticed, it usually shows up in the most expensive way possible.
What the dashboard shows and what the business earns are not the same thing
When the cost per click on a search or display campaign falls, that reads as good news. But a cheap click is very often irrelevant traffic — someone comparing prices, researching, or simply curious, carrying no intent to buy. Without a defined conversion, the platform has no signal to tell these apart; all it can see is clicks and impressions, so that is what it optimises towards. The result is a campaign that drifts towards a lower cost per click while reaching an increasingly irrelevant audience. The budget is not shrinking — it is being spent in the wrong place, and that drift never shows up as an error in the report. It reads as improvement instead.
Decisions get made on feel
When several campaigns, audience segments and pieces of ad copy run at once, only conversion data can explain the difference between them: which campaign brought genuine demand, which audience clicked but never enquired, which message was the one that got the form filled in. Without that data, all that is left is impression — "this ad seems to be getting more likes", "this audience looks more active". Decisions to pause or scale get made on those impressions, and impression rarely matches outcome. The most-clicked or most-liked campaign is frequently the one that brings the least business, because catching attention and creating demand are not the same skill.
The one thing that cannot be recovered: the earlier period
Everything else here can be fixed later; this cannot. When conversion tracking goes in weeks or months after a campaign starts, nothing from before that point can be collected retroactively — measurement only produces data from the moment it starts, it does not run backwards. That means nothing can ever be known about the campaign's opening period: which keyword, which audience or which creative actually worked during that stretch is gone for good. Tracking set up afterwards starts a fresh record from that point on, and any comparison with the earlier period becomes impossible. What is lost here is not access to data but a baseline to compare against — the opening period itself can never be produced again.
"Conversion" does not mean the same thing for every business
Conversion is not a single definition; it depends on the business model. For a law firm it might be a form submission or a phone call; for a shop, adding a product to the basket or completing a purchase; for a clinic, an appointment request; for a locally-based service business, a request for directions. Measurement built before that definition is settled ends up tracking something either far too narrow or far too broad, and both produce misleading results. Not every conversion carries the same weight either: a micro-conversion — a brochure download, a long time spent on a page, a video watched to the end — shows interest but guarantees nothing about a purchase; a macro-conversion — a form submitted, a call made, a purchase completed — represents the actual business. Lumping the two together under one "conversion" label pulls optimisation towards interest rather than towards the business itself.
- A form submission or quote request
- A phone call
- A request for directions
- Adding to basket or completing a purchase
Any one of these can be the real conversion for a given business model; substituting one for another disconnects measurement from what that business actually needs.
Phone enquiries and the blind spot of last-click
For many businesses, the real demand happens on the phone — someone sees the ad and calls directly rather than filling in a form. Unless call tracking is set up separately, that call is invisible on the digital side entirely; the ad dashboard shows a click and a cost, with no trace of that click turning into a call. Call tracking closes that gap by showing a different phone number to visitors arriving from the ad, and tying the call back to its source. Skip that step, and phone-driven demand gets reported as if it never happened, making the campaign look weaker than it really is.
Last-click logic creates a separate blind spot. Most measurement models credit a conversion only to the last ad clicked, yet a person typically encounters a brand several times before deciding — watching a video first, then searching for the brand by name, then finally clicking an ad and filling in a form. A report that only looks at the last click makes the contribution earlier in that journey completely invisible, and can lead to shutting down the very channels that built awareness in the first place, on the grounds that they "don't work".
Numbers that look good in a report and bring no business at all
When no conversion is defined, the dashboard does not stay empty — quite the opposite, it finds plenty to show. Reach, engagement rate, video watch time, time spent on a page: these metrics are always available and can paint a consistently positive picture. None of them are bad metrics in themselves, but none of them, alone, show a return on the work.
- High reach can be achieved with low relevance
- High engagement can happen without a single form being filled in
- A long video watch time guarantees nothing about intent to buy
That picture fills a report and looks fine in a meeting, but it does not answer the question a business owner is actually asking: how many genuine enquiries came in this month.
Conversion tracking is not a technical detail — it is the precondition for advertising being legible at all. Whether the channel is a search-based one like Google Ads or one that speaks to an audience that has already shown interest, like remarketing, every decision made before measurement is in place carries the same problem: nobody knows which number is answering which question. Most of that cost shows up not in the budget but in lost time, and in an earlier period that can never be produced again.