Qaurus is a Canadian go-to-market agency. We reviewed all 16 client engagements in our own case study library against a single test, can we show where the number came from, and five passed. This is what those five show, including one result that looks much better in a slide than it does in the data.
Clients are identified by sector and region rather than name.
What is the difference between impressions and clicks?
Impressions count how often your listing was shown. Clicks count how many people actually came. They move independently, and the gap between them is where agency reporting gets flattering.
On a healthcare staffing platform in Canada, measured in Google Search Console and Ahrefs over six months:
| Metric | Change |
|---|---|
| Search impressions | +157%, to 33,000 |
| Clicks | +38%, to 1,910 |
| Ranking keywords | 19 95 (10 in the top three) |
| Indexed pages | 20 334 |
| Referring domains | 131 256 |
Impressions grew roughly four times faster than clicks. Both numbers are real and both are good. But "search impressions up 157%" is the sentence that ends up on a slide, and it is not the one that describes how many more people arrived.
Does a big visibility increase mean click-through rate improves?
No, in this case it roughly halved. Working from the figures above, click-through rate went from about 10.8% to about 5.8%.
That is not a failure. Ranking for 95 keywords instead of 19 means appearing for many broader, lower-intent searches that were previously out of reach, and those convert to clicks at a lower rate by their nature. A falling CTR alongside rising clicks is what expansion looks like.
It is worth knowing because the reverse also happens: an agency can report improving CTR simply by ranking for fewer, narrower terms. Neither number means anything on its own. Ask for both, plus the absolute click count.
How long does this take?
Where our records state a period, it is six months. Two of the five engagements carry an explicit timeframe in the source, and both are six months, the healthcare staffing platform above, and a fashion e-commerce brand in Pakistan whose sales grew 150% over the same span.
The other three report outcomes without a stated window. We have not inferred one. If an agency tells you how long something took, ask where that figure is recorded.
What do results cost, when they are actually measured?
Two of the five have a stated cost per outcome, and both come from ad-platform reporting:
| Sector, region | Outcome | Measured cost | Window |
|---|---|---|---|
| Payments / fintech, UK | 337 leads via Meta lead forms | £21.70 per lead | 1 Jul 2025, 31 Mar 2026 |
| Consumer app, UK | 208 tracked iOS installs | £1.04 per install | last three months |
The fintech engagement also produced more than 50 qualified leads a month, 15 to 20 of them from organic search, with ranking keywords growing from 998 to about 1,300.
These two are the only engagements where we can put a unit cost on a result. That is worth saying plainly, cost per lead is the number buyers most want and the one agencies least often publish.
How much of the result came from organic search?
Where it is separated, organic did most of the work, but not all of it.
- Industrial manufacturer, Pakistan (Google Analytics 4): 67.7% of all sessions, 7,684 of 11,352, came from organic search, and 66.2% of those organic sessions were engaged.
- Payments / fintech, UK: 15 to 20 of 50+ monthly qualified leads came from organic, so roughly a third, with paid social carrying the rest.
The manufacturer's figure is the cleaner one because GA4 separates the channel directly. The fintech split comes from lead reporting, which attributes more coarsely.
Why we left eleven engagements out
Our case study library holds 16 engagements. Eleven of them carry a headline figure, "3x conversions", "300% traffic", "40% more patients", with no stated measurement period, no named source, and no starting number.
Those results are real work. But a percentage with no baseline and no source is a claim, not evidence, and we have just spent a month building checks that flag exactly that pattern in our own content. Publishing eleven of them in a study about measurement would be the wrong way round.
What we are doing instead: the eleven stay on their own case study pages, where they are presented as summaries rather than as data. As each is re-measured against a named source, it becomes eligible for a study like this one.
What to ask any agency about a case study
- What was the starting number? A percentage without a baseline can describe almost anything.
- Over what period, and where is that recorded? "Six months, in Search Console" is checkable.
- Clicks or impressions? And what happened to the other one.
- What did it cost per outcome? Cost per lead or per install, with the date range.
- Which channel produced it? "Traffic grew" is not the same as "organic traffic grew".
Every one of those questions is answerable from platform reporting the agency already has. An agency that cannot answer them is not necessarily doing bad work, but it is not measuring it.
FAQ
Why anonymise the clients if the data is real? The underlying case studies are published under client names on our own site; the anonymisation here is by sector and region so the comparison reads on the pattern rather than the logo. The measurement source is named in every case, which is the part that makes a figure checkable.
Is a 46% drop in click-through rate a problem? Not on its own. It happened while total clicks rose 38% and ranking keywords went from 19 to 95. Falling CTR with rising clicks usually means a site has started appearing for broader searches. It would be a problem if clicks were flat or falling too.
Do these results predict what another business would see? No. Five engagements across five sectors and three countries is a description of what happened, not a forecast. Different markets, budgets and starting positions produce different outcomes, and we have not run a controlled comparison.
Why does indexed-page count appear in a results table? Because it moved from 20 to 334, and nothing ranks that is not indexed. We published a separate study on how slowly indexing actually happens on a new domain, measured across 95 URLs.
Method and limitations
Figures come from the original client reporting cited in each case study: Google Search Console and Ahrefs, Google Analytics 4, Meta Ads Manager, and client-side lead reporting. Each is named above with its engagement. Derived figures, the click-through-rate change and the impressions-to-clicks ratio, are arithmetic on the stated numbers, and the inputs are shown so the working can be checked.
This is five engagements, not a benchmark. They span healthcare staffing, manufacturing, fintech, a consumer app and fashion e-commerce, across Canada, the UK and Pakistan. The sectors and the measurement tools differ, so the figures are not comparable to one another and no average of them would mean anything. We have deliberately not computed one.
We have not run counterfactuals. None of these engagements had a control, so we can describe what changed during the work but cannot isolate how much of it the work caused.
The eleven excluded engagements are a real gap in our own records, not a judgement on the work. Re-measuring them against named sources is in progress.
We run this measurement on our own site continuously, and on client work where the reporting supports it. If you want the five questions above answered about your own marketing, get in touch.
