Home / Blog / Analytics

Analytics · 4 min · 26 August 2026 · Updated 2026-08-26

Why do Direct and Metrica show different numbers, and which one should you trust?

Short answer

Direct counts clicks, Metrica counts visits. These are two different events, and the discrepancy between them is normal by design, not a breakage. The question is not which one to trust but which number to use for which decision: spend always from Direct, behavior from Metrica, conversions from a single source with a single attribution model. The warning sign is not the gap itself but a sharp change in it.

Artur KupriyanovArtur KupriyanovFounder of Neuron

They don’t count the same thing

A click is counted by Direct the moment a person taps the ad — an event on the ad system’s side, before anything has loaded. A visit is counted by Metrica when the counter has run on the person’s device — an event on your site’s side, after the page has loaded.

A whole life lies between those two moments: the connection, the site’s speed, the person’s patience, their browser settings. The numbers are never identical, and demanding that they be is demanding the impossible.

Seven causes of the gap

1. The person left before the page loaded

The most common one. The click is paid for, the visit didn’t happen. It is especially noticeable on mobile internet and on heavy sites. A large gap here is not an analytics problem but a speed problem, and it costs you money twice.

2. One visit from several clicks

Metrica merges activity into one visit if the person returns within thirty minutes. They left, thought it over, clicked again — Direct will count two clicks, Metrica will keep one visit.

3. Several visits from one click

The reverse case: someone came from an ad, closed the tab, came back directly an hour later — one click, two visits.

4. Blockers and disabled scripts

The Metrica counter is a script, and part of the audience doesn’t run it: ad blockers, corporate policies, data-saving modes. The click is still counted.

5. Filtering of invalid clicks

Direct filters out clicks deemed invalid and doesn’t charge for them, but this doesn’t happen instantly. A report pulled today and the same report a week later may differ.

6. The counter isn’t installed everywhere, or it’s the wrong one

A boring and very common cause: the counter was never installed on the campaign’s landing page, or the site has two counters and you’re looking at the wrong one, or there’s a mistake in the number. It takes a minute to check — open the landing page on your phone and look at the real-time traffic report.

7. The tagging got lost

If the link is set up with tags in URLs rather than the built-in integration, a redirect can strip the tags. The classic case: the ad leads to the address without www, the site redirects to the address with www and loses the tagging along the way. The visits are there, but their source is “direct traffic”.

The numbers never match. What you can have is a clear and stable gap — and that is what to aim for instead of equality.

Which number to use for which decision

  • How much was spent — Direct. Direct is what charges the money; there is no other data.
  • How many clicks were bought — Direct. A click is its own event.
  • How people behave on the site — Metrica. Only Metrica sees behavior.
  • How many leads — one source, chosen in advance. Otherwise the numbers can’t be compared from month to month.
  • Cost per lead — spend from Direct divided by leads from Metrica.

The last item is the only correct way to calculate cost per lead without end-to-end analytics: spend is taken from whoever charged it, results from whoever saw them.

When the gap becomes abnormal

There is no general threshold like “a discrepancy of up to fifteen percent is acceptable”: it depends on the niche, site speed and audience mix. Any universal figure you are given is pulled out of thin air.

The working approach is different. Export two numbers month by month for half a year: clicks from Direct and visits from the ad source in Metrica, and calculate the ratio for each month. You’ll get your own norm — for example, “visits are consistently a fifth lower than clicks”.

From then on, the signal is not the gap itself but a change in it. It used to be steady at one value and became twice as bad — something has happened to the site, the tagging or the traffic, and you should look through the list of seven causes above.

A twenty-minute check

  1. Open the ad’s landing page on your phone. Do you see yourself in Metrica in real time?
  2. How many counters are on the site? Check the source code.
  3. Do the ads lead to the same address variant the site lives on — with or without www, http or https?
  4. In Metrica: do ad visits have a source, or do they fall into “direct traffic”?
  5. Clicks and visits for six months, the ratio for each. This is your norm.
  6. Is the counter linked to the ad account? When they are linked, part of the discrepancy goes away on its own.

What you can’t achieve

Equality. It is impossible to bring clicks and visits to one number with any setup. The goal is a clear and stable gap.

A single truth about conversions. An attribution model is a choice of how to split credit between sources, and different models give different numbers on the same data. There is no “correct” model; there is one chosen in advance and not changed in the middle of a comparison.

An answer to the question “where did the client come from”. Analytics sees visits, not deals. Linking an inquiry to its source requires separate steps, and they can be done for free — covered separately.

Where to go next

The answer above is based not on general opinion but on measurements we take ourselves.

More in this section

To answer your question specifically

The article answers in general; a review answers about your site: a 100–150 page crawl, rankings, demand and a comparison with the top results. An hour of our time, free of charge.

Write to us