Rankings are marketing: the argument, made properly

The ranking table is the most trusted document in management education and the least examined. It looks like the reading from an instrument — as if someone held a ruler against two hundred institutions and wrote down what the ruler said. It is nothing of the kind. It is an editorial product, assembled by a publisher, and every choice that assembles it is as arguable as anything else an editor does.

That is not a conspiracy claim. It is a description of how the thing is made, and it can be checked against any ranking’s own methodology page — the page almost nobody reads.

What a ranking actually is

To build one you make three decisions. Which criteria count. How much weight each carries. Whose answers you trust. None of the three has a correct answer. Give research output a heavier weight and the research universities rise; favour entry salaries and the metro campuses rise; nudge the weights a little and the table reorders itself without a single institution having changed that year. Fourteenth place against nineteenth is not a fact about teaching. It is a fact about a spreadsheet’s settings.

Then there is the data, much of it supplied by the very institutions being ranked, with every incentive to present generously and thin verification behind it. We have written before about placement numbers that cannot be checked; those same numbers flow into ranking submissions, and the published table lends them a rigour they never earned.

The reputational components are stranger still. Surveys poll deans and recruiters — people scoring institutions they have not attended and mostly never visited. What such a survey retrieves is the memory of earlier rankings, which is how a list becomes self-confirming while presenting itself as fresh evidence.

Who the customer is

The reader pays nothing for a ranking, which is always the moment to ask who does. Institutions buy advertising around the lists, license badges for their prospectuses, and sponsor the summits at which the lists are unveiled. The publisher’s incentive runs one layer further down: a stable table is not news, and a list that is not news cannot be sold again next year. Movement is an editorial requirement of the product, not a finding about the schools.

A table on which nothing moves is a table that cannot be published twice. Churn is a requirement of the product.

The narrow use that survives

Made properly, the argument is not “ignore rankings”. They perform one real service: they compress an unsearchable field into a list of names, and the difference between the top of a list and the bottom is usually genuine. The precision fails exactly where it is most often cited — between neighbours. Nobody can defend the proposition that the eleventh programme is better than the fourteenth, because the proposition was never about the programmes. It was about the weights.

And there is a whole category of question a ranking cannot see at all: whether the teaching this year is any good, how a programme treats a student who falls behind, what happens to the ordinary graduate rather than the decorated one. For an online or distance MBA the mismatch is worse, because much of what the criteria measure — campus life, research standing, a full-time placement machine — is precisely the part you will never touch.

How to read one anyway

Use a ranking the way you would use a stranger’s shortlist: as names to investigate, never as a verdict. For each name, ask the questions the table skipped. What will this cost in total, in writing? Which document proves the award is entitled? What would my actual week look like? Recognition — the thing that genuinely protects you — is binary and appears in no weighting anywhere; we have argued that above a floor the extra money buys teaching, not standing, and no table captures that distinction either.

Our own answer to the ranking problem is structural: we grade programme types, not named colleges, because a judgement about routes can be defended in public and three thousand judgements about names cannot. The rest of what this blog holds is laid out, position by position, at our start-here page — written to be disagreed with, which is more than a weighting can say for itself.