We Checked Which Search Engines Actually Send Us Traffic. Bing Beat Google.

We optimise for Google. Everyone does. So it was uncomfortable to pull 90 days of session data and find that Bing sends us slightly more organic traffic than Google does — and that Google Search Console, the instrument our entire reporting stack rests on, can see about 1.7% of our sessions.

Organic sessions by source, 90 days

14 May to 11 August 2026, across 21 sites, from GA4. Bing is ahead of Google by four sessions, which is a rounding error dressed as a finding — the honest statement is that they are level, not that Bing won.

We would not generalise this to your site and you should not either. These are small utility and calculator sites with essentially no backlink profile, which is precisely the profile where Google is hardest and Bing is comparatively winnable. A site with real authority would probably see the reverse. What travels here is the habit of checking, not our number.

The part that actually matters: the window changes the answer

Same property, same metric, fourteen points apart. Neither is wrong. They answer different questions, and at n = 230 a fortnight of variance moves the share a long way.

We are not describing this as a hypothetical. Three days before writing this, our own internal status file recorded that “Search Console observes roughly a fifth of our organic traffic.” That was the 28-day window, stated as though it were a property of our traffic rather than a property of our window. Over 90 days it is more than a third. We corrected the record and then wrote this, because a measurement mistake we caught ourselves is worth more published than buried.

What Search Console can see

Search Console reports Google Search and nothing else. That is not a flaw — it is the product. It does become a problem when it is used as a general dashboard for “organic performance,” which is roughly what we had been doing.

AI assistants sent us more traffic than every search engine combined. 265 against 230. That comparison is like-for-like — same property, same window, same measurement — so we trust its direction more than either absolute number.

The 82.7% we cannot explain

Direct at 4,085 sessions is far too high to take at face value, and we would rather flag it than build an argument on top of it. Some is genuine repeat use of tools people bookmark. A large share is almost certainly misattributed: assistant traffic, app-to-browser handoffs, and referrer stripping all land in Direct.

So treat every “share of total sessions” figure above as carrying that uncertainty. The organic-source split does not, because it compares organic sources against each other and any misattribution hits them alike.

What we would suggest doing

• Look up your own organic source breakdown. Two minutes in GA4. If you have been assuming “organic” means Google, it might not.
• Report the window with the share, and check a second window before you believe it. Our fourteen-point swing came from doing the first thing and skipping the second.
• Do not stop using Search Console. It is excellent at what it measures. Just stop reading it as though it measured the whole of search.

We have written before about mistaking a broken metric for a broken business. This is the same failure with the polarity reversed: not a metric that lied, but a metric that told the truth about a smaller thing than we thought we were asking about.

Frequently asked questions