Review · Banking · 23 Oct 2026
Brand or non-brand: what is dragging a bank’s organic traffic down Two pictures in one dataset: over a year brand falls, over five months non-brand catches up
Over a year a bank’s brand traffic fell 38% and non-brand 18%, and it is tempting to conclude “awareness is suffering”. But within January–May non-brand falls faster in both engines. Here is how the two pictures fit together and what to check in each.
- Yandex, brand157 → 109k−31%
- Yandex, topical277 → 182k−34%
- Google, brand326 → 266k−18%
- Google, topical49.6 → 32.8k−34%
Internal data of a large bank, web analytics and a rank tracker, 2025–2026. Values are anonymised: absolute figures are altered and rounded, and relative changes are recomputed from the numbers shown. Shares are our calculation from the values in the table.
In short
- Over a year brand fell more than non-brand (−38% vs −18%). From January to May non-brand falls faster in Yandex (−34% vs −31%) and in Google (−34% vs −18%).
- The two pictures are built on different bases (year over year and five months), so they cannot be interpreted together; check each separately first. This is not a conclusion about “search in general”: the two pictures may have different mechanisms behind them, but the data do not prove it.
- A brand decline should not be pinned on SEO: it combines demand for the name, the shape of the results page for brand queries, and attention to the brand outside search. SEO affects only the second.
Two pictures: a year and five months
Year over year (May to May) brand loses twice as much as non-brand. Over five months (January → May) the picture reverses. These are different horizons and different bases (see the note below), so they cannot be directly compared.
| Horizon | Brand | Non-brand | Falling faster |
|---|---|---|---|
| May 2025 → May 2026 | 1.02M → 632k (−38%) | 448 → 368k (−18%) | Brand |
| January → May 2026, Yandex | 157 → 109k (−31%) | 277 → 182k (−34%) | Non-brand, slightly |
| January → May 2026, Google | 326 → 266k (−18%) | 49.6 → 32.8k (−34%) | Non-brand, twice as fast |
To say when and what fell, a monthly 2025 series of brand and non-brand on one base is needed: the data do not have it. So we do not claim that brand fell before non-brand.
The brand/topical splits and the channel summary are built on different bases, and their sums do not match. Compare percentages inside each and do not add them up; see the main research.
Audience structure: Google is almost entirely brand
In Google the brand share in this data is 87% in January and 89% in May; in Yandex 36% and 37%. So “Google falls less” is about the query mix, not search engine quality: in Yandex both types fall nearly equally, in Google brand gives the volume and non-brand the rate of decline.
A consequence for reporting: engines cannot be compared by total trend without a brand/non-brand split. The share table and more are in Google vs Yandex: volume and conversion.
Where a brand decline comes from: three mechanisms
A brand query comes from someone who already knows where they are going. For them not to arrive, one of three things must change.
- Demand for the name. Fewer people search the bank by name. For two products (savings and settlement accounts) brand queries in Wordstat fell 41–42% in May against April, while base demand fell 22–23%. This is a single-period example (May vs April, two products): it explains neither the annual −38% nor the five-month dynamics.
- The results page for brand queries. Answer blocks, competitor ads and service widgets take clicks even when demand and position are unchanged. It shows in Search Console impressions and CTR, not in traffic.
- A different route. People go to the app or a bookmark. Then Direct would grow, but over five months it fell 6% (1.51M → 1.42M), so these data hold no direct sign of such a shift.
The first mechanism is not about SEO. The second is about SEO and the results page. The third is checked through neighbouring channels: paid and direct.
Brand is falling: a six-question check
Answer “yes”, “no” or “don’t know” about your site for the same period.
- Is the list of brand queries fixed once, including spelling variants, transliteration and “product + brand” forms?Brand definition
- Is demand for the name (Wordstat, Trends) checked against brand traffic on the same sample?Demand
- Do brand-query impressions in Search Console fall together with clicks?If impressions hold, it is the results page
- Have you reviewed results for 10–20 brand queries by hand for new answer blocks and competitor ads?Manual review
- Did Direct and the app not grow over the same period?No shift
- Is non-brand judged on a fixed set of pages rather than the whole section?One sample
This review’s own working scale, not an industry standard. Any “don’t know” answer makes the page show “Gaps remain” and no verdict.
What to do about brand and non-brand declines
- Fix the list of brand queries and keep it with the report.
- Show brand and non-brand as two lines over two horizons: a year and the last 3–5 months.
- Do not set a traffic-growth KPI for brand; set one on the click share of brand impressions.
- Break non-brand down by page type: commercial, reference, definitions; the causes differ (section breakdown).
- Look at brand results with your own eyes once a month: new blocks and ads are visible at once.
FAQ
Is a brand decline an SEO problem?
Not necessarily. Demand for the name depends on awareness. SEO affects what is shown for a brand query and how many clicks remain at the same demand.
How can a demand drop be told from a click drop?
By impressions in Search Console for brand queries: if impressions fall, demand changed; if impressions hold while clicks fall, the results page changed. More on impressions and position in the guide on position vs clicks.
Why does non-brand in Google fall twice as fast as brand?
Non-brand is small in Google (11% in May), so its percentages are noisier. The trend exists, but comparing percentages with Yandex needs care.
What should a client see?
Two lines over two horizons, the brand share by engine, and a note that SEO does not control demand for the name.
Sources and method
- Search Console Help: impressions, position and clicks — how to tell impressions from clicks
- Internal data of a large bank: web analytics, rank tracker, Wordstat (2025–2026)
Method: data of one large bank, January–May 2026 and May to May. Values are anonymised: absolute figures are altered and rounded, and relative changes are recomputed from the numbers shown; brand shares are our calculation from the values in the article. The data hold no monthly 2025 series, so the point where brand began to fall is not determined. As of 23 Oct 2026.