Koolav next to Clutch, G2 and DesignRush: what is actually different
Checked against what each of them publishes about itself, on 28 August 2026. The difference is smaller than we expected on money, and larger than we expected on negatives.
The claim we did not make
The easy version of this page says the directories are pay-to-play and Koolav is not. We went to read their own terms before writing it, and that version does not survive. G2 states that sellers may not pay to influence ratings, placement or sort order. Clutch sells sponsorship and says plainly that sponsors appear above non-sponsors on directory pages — while also stating that sponsorship does not move its underlying Clutch Rank, which is a real distinction and we are not going to blur it. DesignRush is the clear case in the other direction: it sells sponsorship from $500 to $5,000 a month and describes the benefit, in its own words, as a top position in a category ranking.
So money buys visible position in some places and not others, and a buyer genuinely cannot infer much from that alone.
The difference that does hold
None of them publishes a negative. Every directory in the table below is a list of participants. A firm that would fail an evaluation is not listed with a bad verdict — it is absent, and absence is indistinguishable from never having joined. Koolav publishes 35 cuts out of 734 evaluations, by name, with the reason and the sources, and they sit on the same pages as the recommendations.
That is the whole asymmetry. A list that can only add is a list whose silence carries no information. The cuts are what make the rest of the library worth reading, which is why they are not filed somewhere quieter.
Point by point
| The question | Koolav | The directories |
|---|---|---|
| Does it publish the ones it rejects? | Yes. 35 of 734 evaluations are CUT (4.8%), named, with the reason and the sources. | No directory here publishes a negative verdict. A firm that would not qualify is absent rather than listed badly, and a buyer cannot tell that apart from a firm that never joined. G2 documents the reasons a product may not qualify for a listing; the outcome is exclusion, not a published finding. source |
| Can a company pay for position? | No. Koolav takes no payment from any agency for a listing, a placement, or a verdict, and no attribute a company can buy enters a score or an ordering. | It varies, and the honest answer is not one answer. DesignRush sells sponsorship from $500 to $5,000 a month and states the benefit as a top position in a category ranking. Clutch sells sponsorship and says sponsors are listed above non-sponsors on directory pages, while stating that sponsorship does not affect its Clutch Rank score. G2 states that sellers may not pay to influence ratings, placement or sort order. source |
| Is there a published rubric, per discipline? | Yes. One rubric per category, 32 of them, with weighted dimensions shown on every evaluation next to the score they produced. | We located no published per-discipline rubric on any of them. What they publish is one ranking model applied to the whole company — reviews, market presence, profile completeness — so the same score answers "is this firm well regarded" and never "is this firm right for the job I am buying". If a discipline-level standard exists somewhere on these sites, we did not find it, and we are stating that rather than inferring it does not exist. |
| Where does the evidence come from? | Public sources we read ourselves. 3,607 verified findings and 5,076 citations across the library, each one a link you can open. | Client reviews, most of them solicited by the agency being reviewed. That is real evidence about past engagements and we do not dismiss it — but it is testimony gathered by the interested party, and it cannot tell you whether a firm publishes its pricing, who owns the work, or whether a named client is still a client. |
| Does it say what it does not know? | Yes. 3,212 open questions are published across the library, on the same pages as the verdicts. | Not established on any of them. We found no directory that publishes what it could not verify about a listed company. |
| Is every evaluation dated? | Yes, all 734 of 734. The date is on the page, and an old one is shown rather than hidden. | Profiles carry review dates. The ranking itself is continuously recomputed and generally not dated, so a list captured today and a list captured last year are not distinguishable after the fact. |
| Can the company answer back? | Yes, and the answer is published either way. Every evaluation carries a right of reply and there is no paid path to a different verdict. | Yes on the review platforms — vendors can respond to individual reviews. There is no equivalent mechanism for disputing a ranking, because no ranking states a finding specific enough to dispute. |
| Is the score about fit, or about popularity? | Fit for a described job. Scores are comparable within a discipline and deliberately not across them, and several records say outright that a buyer shopping for something else should ignore them. | Largely accumulated standing — review volume, recency, market presence. Those measure how much attention a firm has gathered, which correlates with size more than with suitability. |
What the directories are better at
Coverage, by a wide margin. Clutch and G2 hold vastly more companies than this library does, and if you need to know whether a firm exists and roughly what it sells, they will answer faster. They also carry something we do not have and cannot manufacture: volume of first-hand client testimony. A hundred reviews of an engagement is evidence about that engagement, and no amount of reading a website replaces it.
What that testimony cannot tell you is whether the firm publishes its pricing, who owns the work when the contract ends, whether a named client is still a client, or whether the case study is four years old. Those are the questions this library exists to answer, and they are answerable from public evidence without asking anyone's permission.
Check this page
Every claim about another company above is linked to that company's own page. The Koolav numbers are computed from the library when this page is requested, not written into it — so they cannot go stale, and if the cut rate falls this sentence gets weaker on its own.