Should I drop my cheapest tier?

Cerno is a private AI boardroom that runs consequential decisions through a structured process and returns a brief like the one below. This example is illustrative; your own would reflect what the tier actually earns, who's on it and what they do next.

How this brief was reached

The advisors formed independent positions, a Devil's Advocate challenged the leading one, and the vote landed 4–1 for closing the tier — but not for removing it, which is the distinction that the whole brief turns on.

The Customer Advocate's dissent is the reason. The cheapest tier doesn't earn much in its own right, and that's the first-order view. The second-order question is what it does for everything else: it's the way in for customers who upgrade later, and it's the reference point that makes the middle tier look reasonable. Neither of those shows up in the tier's own margin, and neither is easy to see until the tier is gone — at which point the customers who would have come in through it simply don't arrive, and you never see them.

That's why the recommendation is staged. Closing the tier to new customers, with existing ones grandfathered, converts an irreversible removal into a reversible test. If new-customer numbers hold up on the middle tier, the on-ramp wasn't needed and the tier can be removed with evidence. If they fall, it can be reopened having cost a quarter.

Confidence is moderate. The board is confident the tier is a support burden and a weak earner; it is less sure the on-ramp effect is as small as the margin suggests, which is exactly what the quarter tests.

What the board weighed

Count and revenue point in opposite directions. Most customers are on the tier; little revenue comes from it. Read one way, it's a distraction. Read the other, it's the largest group of people who've chosen you, and the ones with the most room to grow. The board declined to pick a reading and designed a test instead.

Support cost may be a delivery problem, not a tier problem. The tier generates disproportionate support. The board asked whether that's because of who buys it or because of how it's served — no onboarding, no self-service, a gap in the documentation. If the second, the cost can be cut without cutting the tier, and that's cheaper than losing the on-ramp.

Changing what you sell is not changing what people bought. The single biggest source of pricing anger is customers discovering that something they already pay for has changed underneath them. Grandfathering is the whole difference. The board treated it as non-negotiable rather than as a nicety.

A quarter is the honest length. Shorter and one good or bad month decides it. Longer and, if the tier was doing a job, the cost of the test compounds. One quarter, compared against the previous one, on total new customers rather than middle-tier conversion alone — because the question is whether people arrive, not which door they use.

If your situation differs

If upgrades from the tier are rare — well under one in ten — the on-ramp argument weakens and the Customer Advocate's dissent carries less weight. Close it, test for a quarter anyway, and expect to remove it.

If the middle tier is a big price jump from the cheapest, the test is likely to fail: the people who'd have bought cheap won't buy middle. Consider whether the answer is a restructured entry tier — fewer features, same price — rather than none.

If the tier's customers are a different kind of customer entirely — hobbyists on a product built for businesses, say — this is a positioning decision more than a pricing one, and it belongs with the rebrand brief's question about who the business is for.

The frameworks behind it

This decision turns on second-order thinking — what the cheapest tier does for everything above it — and on reversibility: closing rather than removing is the staged move that keeps a one-way door two-way while the numbers come in, the same logic as the price-rise brief.