What suppliers say about themselves, and what it is worth

How should a capability list be read?

Subtractively. An assertion is close to uninformative, since the incentives push everybody towards the identical set; an omission had to be a deliberate choice.

Consider two entries. One lists nine capabilities spanning consulting, monitoring, testing and compliance. The other lists two. The nine-item list is the safe commercial choice and tells you almost nothing. The two-item list is a firm that accepted being invisible to most searches in exchange for being right for a few, and that decision is a real signal about how they work.

The second useful reading is order. People put first what they most want to be hired for, and the last item on a long list is frequently something the firm can do rather than something it does. Reading a list backwards is a reasonable heuristic for finding the parts that were added defensively.

The third is vocabulary. An account phrased in the language of outcomes — what the client ends up able to do — usually comes from somebody who has delivered the work. A description written in the language of technologies is often written by marketing, and the distinction survives contact with the first meeting more reliably than any credential.

A caution against reading too much into any of this. These are heuristics for allocating attention across a long list, not verdicts. A superb practice may carry a dreadful summary because whoever drafted it was busy delivering, and a polished paragraph may belong to a firm that hired an agency. The value of the exercise is deciding which eight of two hundred entries deserve an hour, and that is genuinely all it is for.

One further reading rewards the patient. Compare an entry against the website it points at. Where the two agree closely, somebody maintains both and the operation is tidy. Where the site describes a different business entirely — a narrower niche, a broader platform, a pivot — you have learned something no single document could have told you, namely the direction the company moved after writing that sentence.

Is a smaller supplier riskier?

Differently risky, and the honest framing is a choice between two failure modes rather than a ranking.

A large incumbent will not disappear. What it will do is fold your product into a platform, move its pricing into a bundle, and assign you an account manager whose other clients are larger than you. Your requests enter a roadmap process where they compete with everybody else's, and the probability that a specific thing you need gets built is low regardless of how reasonable it is.

A small outfit gives you attention that is real. The people who built the product will answer your questions, changes you ask for may appear in weeks, and the relationship is with people rather than with a process. What you are carrying is the possibility of acquisition, a support function that is three people, and a roadmap that depends on the company continuing to exist in its present form.

Neither risk is avoidable and both are manageable, but by different clauses. Against the large supplier, protect the ability to leave: data export, detections expressed outside the product, no engineered exit penalty. Against the small one, protect continuity: an escrow arrangement if the product is central, and terms that survive a change of ownership.

The mistake worth avoiding is choosing on size as though it were a quality. It is a shape of risk, and the useful question is which shape your contract is written to survive.

There is a middle case that gets overlooked because it has no marketing behind it. An outfit of thirty people that has stayed thirty for a decade is neither the growth company heading for an exit nor the platform vendor with a roadmap process. It chose to remain a size at which the founders still do the work, and it is frequently the best available answer for a narrow problem — while being invisible in every comparison built around scale.

Should you buy through a reseller?

It depends entirely on whether the intermediary is doing work or forwarding messages, and the difference is testable before signing anything.

A reseller earns their margin in two situations. The first is when they genuinely operate the product for you, in which case you are buying a service and the underlying licence is an implementation detail. The second is when they aggregate support across several products you own, so one relationship covers a problem that would otherwise involve three vendors blaming each other.

They do not earn it as a purchasing intermediary. If their contribution is forwarding your technical questions to the vendor and forwarding the answers back, you have added a delay and a margin to every interaction you will have for the life of the contract.

The test is a single technical question at the first meeting — something specific about how the product behaves in a situation you care about. A partner who operates the thing answers it. A partner who resells it says they will find out, and that answer is the whole evaluation.

Where an intermediary genuinely adds value and it is worth saying, because this section reads harshly otherwise: for an organisation without the staff to run a product properly, a competent operator is frequently better value than the licence plus an unfilled role. The question is what you are buying, not whether buying it is legitimate.

Where to start on a Monday

Three things, none of which requires contacting anybody.

Write the problem down in one sentence, in your own words. Not the category you think it belongs to. If the sentence contains a product category rather than a situation, it was written after the shortlist rather than before it.

Trace what became of the last three vendors you bought from. Still independent? Same owner? Same product name? Whatever base rate emerges applies equally to whoever you choose next, and it is usually worse than people assume.

Write the failure condition for the pilot before the pilot. One sentence naming what would have to be true, by when, for the answer to be no. A pilot without one concludes with everybody having been impressed and no decision available.

After those three, the harder work is prioritisable: getting exit terms into contracts at signature rather than renewal, deciding which shape of vendor exposure you are equipped to carry, and building the habit of describing problems rather than categories. An organisation that has done the first three can state what it needs in a sentence, which turns out to be the scarce thing in the whole transaction.

A closing observation about sequence. Most purchasing conversations begin with a budget and end with a requirement, which is backwards and produces a specification assembled to justify a figure. Beginning with the sentence, then discovering what addressing it costs, sometimes yields an uncomfortable answer — and an uncomfortable answer arrived at honestly is still a better basis for a decision than a comfortable one reverse-engineered from a number somebody had already approved.