Reviews

We do not collect reviews

Our own research page says we exclude review aggregates from our comparisons because incentivised reviews are widespread enough in this category that the averages do not mean what they appear to mean. It would be incoherent to publish that and then solicit reviews of ourselves.

Want to speak to a customer?

Tell us your size and industry. We will connect you to a comparable customer, including one where it went badly.

1 / 3
No incentivised reviewsNo paid review placementsReferences offered instead

Reviews

Four things wrong with software review aggregates.

None of these are hypothetical, and all four are visible on the major platforms in this category if you look for them.

Reviews are routinely incentivised

Gift cards, account credits, and prize draws in exchange for a review. Platforms disclose that a review was incentivised; almost nobody filters by it, and the effect on the average is not disclosed.

They are collected at the wrong moment

Vendors solicit reviews shortly after go-live, when relief is highest. The opinion that matters — eighteen months in, after an upgrade and an audit — is systematically under-represented.

Placement is frequently purchased

Category positions, comparison placements, and "leader" badges on several major platforms are commercial products. The badge and the score come from different mechanisms and look identical.

Unhappy customers rarely post

A company mid-way through a failed implementation is dealing with the failure, not writing a review. The distribution is missing exactly the observations a buyer most needs.

What we offer instead

References, including difficult ones. We will connect you to a customer of comparable size and shape, and — if you ask — to an engagement that went badly. We have had migrations where the reconciliation clock restarted twice and pushed a cutover by a quarter, and a rescue we could not save.

Measured outcomes with the distribution shown. Close length, straight-through rates, and days sales outstanding, measured from production data rather than reported in a testimonial, with the interquartile range published rather than the best case.

An evaluation on your own data. Send several hundred historical transactions with known outcomes and we will score against them before you commit, and show you where it fails. That predicts your experience in a way no review can.

A review tells you how somebody else felt. A test on your own transactions tells you what you would actually get.

If you are reading reviews anyway

You should — they are not worthless, they are just weaker evidence than they appear. Some ways to read them better:

  • Filter to reviews over twelve months old where the platform allows it. The honeymoon has passed and an upgrade cycle has happened.
  • Read the two- and three-star reviews first. They are usually the most specific and the least incentivised. Five-star reviews with no detail are close to noise.
  • Check whether reviews are marked as incentivised, and whether the platform lets you exclude them. Compare the average with and without.
  • Look for implementation detail rather than sentiment. “Great product” is not information. “Our data migration took four months longer than quoted” is.
  • Discount badges entirely unless you know how they were awarded. On several platforms they are purchasable.

What we will never do

  • Offer an incentive of any kind — gift card, discount, credit, or prize entry — in exchange for a review.
  • Purchase category placement, comparison positioning, or a badge on any review platform.
  • Solicit reviews immediately after go-live, or from customers we know are currently pleased.
  • Publish a score without the sample size and the collection method beside it.

If a customer writes a review unprompted, that is entirely their business and we will not ask them to change it. What we will not do is manufacture a distribution and then present its average as evidence.

The cost of this position

It is real. Buyers filter shortlists by review count and score, and a vendor with forty reviews looks more established than one with four regardless of what the reviews say. We lose consideration that way.

We think that is the correct trade given that we publish a page criticising review aggregates in our comparison methodology. Holding a position only when it is free is not holding a position.

This applies to our directory too

We exclude review aggregates from our own scored comparisons of other vendors, for the same reasons set out above. That is stated on our methodology page, and it is why our directory scores products on functional fit, cost, time to value, implementation risk, extensibility, and ecosystem rather than on what reviewers said.

Questions

Common follow-ups.

Why no reviews at all?
Because we exclude review aggregates from our own comparisons of other vendors, on the grounds that incentivised reviews distort the averages. Soliciting our own would be incoherent.
Are reviews worthless?
No, just weaker evidence than they appear. Filter to older reviews, read the two- and three-star ones first, and look for implementation detail rather than sentiment.
What do you offer instead?
References including difficult ones, measured outcomes with the distribution shown, and an evaluation run on your own historical transactions before you commit.
What if a customer reviews you anyway?
That is their business and we will not ask them to change it. What we will not do is manufacture a distribution and present its average as evidence.
Does this cost you deals?
Yes. Buyers filter shortlists by review count and score. Holding a position only when it is free is not holding a position.

Ask for the reference that went badly.

It tells you more than forty five-star reviews collected in the week after go-live.