Every ranking on Venture Harbour has been through the same process. This page is the manual we follow when we sit down to review a new category.

Would we use it in a portfolio company?

Venture Harbour is an automation venture studio. We operate our own portfolio of automation companies. Every tool we write about is tested against a single question:

Would we run this inside one of our own portfolio companies?

If the answer is yes, it’s considered. If it’s no, we say why.

The testing framework

For each tool we include in a ranked review, we work through the same steps:

  1. Real account. We create a real account and actually use the product.

  2. Real data. We import or set up with the same kind of data a real customer would.

  3. End-to-end use. We use the tool to complete the core job the category exists to do — run an email campaign, close a deal, deploy a site, write some code — not just click through the settings.

  4. Compare side-by-side. Every scored tool in a category is tested against the same criteria on the same rubric, by the same reviewer.

The scoring rubric

Each tool gets a rating out of 5 across the dimensions that matter for that category. The weights differ — page builders care about rendering speed; CRMs care about reporting depth — but the consistent ones are:

  • Core job performance. Does it do the main thing the category exists to do, well?
  • Ease of use. How long before a competent operator is productive?
  • Integration surface. Does it connect with the rest of a modern stack, or is it a silo?
  • Pricing transparency. Are the numbers on the pricing page what you’ll actually pay, or are there meters hidden behind sales calls?
  • Reliability. Outages, data loss, support response times — the things you only find out in week three.
  • Support and docs. Do they answer, and is the documentation written for operators rather than marketing?

Scores are public. Rating breakdowns appear inside each review, so you can see exactly where a tool gained or lost points.

Who does the testing

Research is written by the author whose name is on the byline. Testing is done by the author — not by an external agency or freelancer — usually inside a live Venture Harbour venture. Every piece is edited and fact-checked against the live product before it goes live, and scored against the rubric above.

When we revisit a ranking

Categories don’t stand still. We re-test a ranking when any of the following happen:

  • A tool in the top three materially changes pricing, positioning, or feature set.
  • A new entrant looks like it’s credibly in the top five.
  • Our own day-to-day use of the category turns up something we missed.
  • Reader feedback points at a factual error (keep them coming — see contact).

The Last updated date on each review reflects the last time we re-tested the category, not a cosmetic refresh.

Affiliate relationships

Some of the tools we rank have affiliate programmes and some don’t. The ranking is written before we know which is which. If the top-ranked tool in a category has no affiliate programme, it still ranks first. For the full policy on how affiliate relationships are disclosed and how they don’t influence ranking, see our editorial policy.

Test Lab Programme partners

A handful of the tools we test are Test Lab Programme partners. They pay us a fee, and it’s worth being precise about what that fee buys — because the obvious assumption is the wrong one.

It buys our time, not our opinion. Testing a platform properly takes days: real accounts, real data, running the thing end to end, then writing it up. The fee covers that work. It doesn’t buy a score, a position in a ranking, a slot in the comparison table, or a kind verdict. Partners get ranked below non-partners regularly, and unfavourable findings go out as we find them. If any of that were for sale the rankings would be worthless — including to the partners paying for them.

Every partner is labelled. There’s a Test Lab Partner (Paid) badge next to their name wherever they appear, a disclosure at the top of any article they’re in, and rel="sponsored" on every link to them.

Where partners do get a say: facts. For 30 days after we publish, a partner can propose corrections to anything we’ve written about them. Where the change is factual or objective — a feature we recorded as missing that isn’t, a price, a plan limit, an integration, a certification — we make it. That’s an obligation, not a favour. Getting a fact wrong is our mistake, and it should be fixed whoever spots it.

Where they don’t: everything else. Our assessment, our emphasis, our score, our ranking and our verdict aren’t up for discussion. A partner can tell us they disagree, and occasionally the argument is good enough that we go back and re-test. What they can’t do is have a conclusion changed because they pay us.

The honest catch is that a paid relationship still buys attention: a partner gets tested, and a tool nobody has paid us to look at might sit in the queue longer. That’s the part worth watching us on. It’s also why non-partners are scored on the identical rubric, and why most of the tools ranked on this site have never paid us a penny.

If a partner thinks a write-up was thin or unfair, they can have the year’s fee back and we take the write-up down. That’s clause 8 of the programme terms, and it’s the only lever they have. Keeping the placement and changing the verdict isn’t on the menu.

Push back

If you think we got a ranking wrong, or a tool is missing from a review that should be in it, let us know.