The conflict of interest, first
BizConsulting sells marketing advisory work. In every niche we rate, we touch the same market — and scoring ourselves alongside people we overlap with would not be honest.
So the rule is simple: we do not appear in any rating on this site. Not at the top, not at the bottom, not with an asterisk. If that ever changes, this page changes first and says so at the top.
A position cannot be bought, accelerated or removed. If paid placement ever appears, it will be labelled as advertising, sit outside the table, and carry no weight in the score. Today there is none — see disclosure.
The six criteria
| Criterion | Weight | What we look at |
|---|---|---|
| Verifiable standing | 25% | Whatever can be checked against an outside source rather than the company's own say-so. The source differs by niche — a directory, an app store, bylined publications. |
| Transparency of terms | 20% | Published fee model, minimum budget, what the fee covers. No pricing is not a zero, but it is not a plus either. |
| Evidence base | 20% | Case studies naming the client, the period and the numbers. "Grew sales 300%" without a client or a period does not count. |
| Named team | 15% | Whether you can see who will actually run the account: names, roles, profiles. "A team of certified specialists" is not a team. |
| Ownership of access and work product | 10% | What stays with the client when the relationship ends: accounts, data, code, content. The cost of switching providers. |
| External corroboration | 10% | Reviews with counts and dates on platforms we do not control, and mentions in independent sources. |
The weights reflect how well a criterion predicts trouble, not how important it sounds. Verifiable standing carries the most weight not because a partner badge makes an agency better, but because a claimed status that cannot be found in the directory is already a signal.
One criterion deserves a separate note, because other ratings omit it: who owns the access and the work product. The question is always the same — what remains with the client when they leave. In paid search that is the ad account with its campaign history and audiences; in organic search it is Search Console, the content and the links; in development it is the code and the store developer account. Almost nobody publishes this, so we ask directly and record the answer. No answer is also a result.
What each criterion means per niche
Two of the six cannot be described identically everywhere. Verifiable standing in paid search means the Google Partners directory; in organic search no such directory exists at all; in development it means the app stores. The weights do not change — what changes is what we go and check.
Google Ads management
| Criterion | What we check in this niche |
|---|---|
| Verifiable standing (25%) | Partner status in the public Google Partners directory. The weight is not in the badge but in the gap: a status claimed on the site and absent from the directory answers the question by itself. |
| Ownership of access and work product (10%) | Whether the ad account stays with the client after the relationship ends, along with campaign history, audiences and years of data. That is the switching cost, not a detail of the terms. |
SEO agencies with a GEO practice
The niche is named this way on purpose. Firms for which GEO is the actual specialisation barely exist in the Ukrainian market yet — the results are SEO agencies that added a service page, and blog posts outranking commercial pages. A list of «best GEO agencies» would rank who published a landing page first, not who specialises.
| Criterion | What we check in this niche |
|---|---|
| Verifiable standing (25%) | There is no directory to check against here. In its place: verifiable authorship — talks, bylined pieces in trade publications, open data from their own research. Claims of being first in the country score zero without something to back them. |
| Evidence base (20%) | For the GEO half, evidence is not rankings but recorded citations: dated screenshots of AI answers, a described measurement method, repeatability. Promising to get you into ChatGPT without a measurement method is not evidence. |
| Ownership of access and work product (10%) | Who owns the access and the work product: Search Console and GA4 accounts, content, acquired links, technical changes. Same question as with an ad account — what remains with the client when they leave. |
Mobile app development
| Criterion | What we check in this niche |
|---|---|
| Verifiable standing (25%) | What gets checked is not a badge but the stores: whether the apps a company claims are actually in the App Store and Google Play, under whose developer account they are published, and when they were last updated. |
| Ownership of access and work product (10%) | Who owns the code, the repository and the store developer accounts once the project ends. An app published under the contractor account is the same lock-in as an ad account held by the agency. |
How the score is built
Each criterion is scored 0 to 100 and multiplied by its weight. The total is the sum, also 0 to 100.
A score below 100 does not mean a weak agency. It means part of the signal is not public: plenty of strong teams simply cannot publish named case studies because of NDAs. That is why every entry carries a "worth weighing" note alongside the score — context rather than criticism.
We do not publish a score at all unless at least four of the six criteria could be verified. The company stays on the candidate list without one.
What we deliberately do not measure
- Client results. We have no access to other people's ad accounts. Any rating claiming to have compared agency performance has compared their marketing materials.
- Price itself. A cheaper agency is neither worse nor better. The score reflects whether terms are published, not what they are.
- Team size. Forty people on staff says nothing about the one who will run your account.
- Industry awards. Mostly paid or reciprocal.
Updates
Data is reviewed quarterly. The current review is 2026-09-17, the next is scheduled for 2026-12-17. Each rating page shows its own last-checked date; if that date is older than a quarter, treat the scores as indicative.
Spotted something wrong about your company? Write to us and we will correct it and date the correction. We cannot change a score on request; we are obliged to fix a wrong fact.