How to Tell a Bad Quarter From a Broken Strategy

Every business initiative reaches a point where you can't tell whether it needs more time or needs abandoning. Persistence and stubbornness feel identical from the inside, and both come with a story about why the results haven't arrived yet.

The way out isn't better judgement in the moment. It's deciding in advance what evidence would change your mind.

Write the falsifier before you start

When you commit to something — a new channel, a service line, a market — write down what would tell you it isn't working, and by when.

Launching the retainer offer. By 30 November I expect: 20 conversations, 8 proposals, 3 signed. If I have fewer than 4 proposals by 31 October, the offer isn't landing and I stop and rework it rather than pushing harder.

That paragraph is worth more than any amount of later deliberation, because it was written when you had nothing invested. The same judgement made in November — after the effort, the sunk cost, and the story you've told people — is a rationalisation wearing the clothes of a decision.

If you can't state what would falsify it, you haven't defined it well enough to test.

Watch leading indicators, not revenue

Revenue is a lagging indicator. It moves a full sales cycle after the thing that caused it, which means waiting for revenue to tell you something puts the decision at least six weeks late.

Watch the upstream numbers:

| Testing | Leading indicator | Lagging | |---|---|---| | A new channel | Conversations generated | Revenue | | A new offer | Proposal rate from conversations | Deals closed | | A price rise | Close rate | Monthly revenue | | A service line | Repeat enquiries, referrals | Contribution | | Marketing generally | Qualified leads per week | Pipeline value |

If conversations are happening but nobody proposes, the offer is wrong. If proposals go out and nobody signs, the price or the positioning is wrong. If nothing generates conversations at all, the channel or the targeting is wrong.

Each of those is a different fix, and the revenue number alone can't distinguish them.

Three failure modes, three responses

The diagnostic that matters, because they look identical and need opposite responses.

The idea is wrong. Nobody wants this at any price. The signal is that even people who should be ideal don't engage — no curiosity, no questions, polite dismissal. More effort makes this worse.

The execution is wrong. People want it; you're delivering or selling it badly. The signal is genuine interest that doesn't convert, or converts and then churns. This is fixable and frequently misdiagnosed as the first.

It needs more time. The signal is that leading indicators are moving in the right direction, just slower than hoped. Proposals rising, conversations improving, referrals starting. Slow but improving is not the same as flat.

The honest test: is the trend flat, or slow? Flat after a fair test means stop. Slow means continue and reassess at the next checkpoint.

Set the decision date at the start

Without a date, you decide when you get tired — and exhaustion arrives on a schedule unconnected to the evidence.

Pick a review date when you commit, tied to something real:

• One full sales cycle for anything sales-related
• Ninety days for a channel — most need that to produce signal
• Three delivery cycles for a new service line

Put it in the calendar. At that date, look at the leading indicators against what you wrote down, and decide. Not "how does it feel" — what did the numbers do against the number you predicted.

Give it a fair test first

The opposite failure, and equally common: changing course every six weeks.

A business that abandons each initiative before it produces signal generates a lot of activity and no compounding. Nothing gets far enough to work.

A fair test needs:

• Enough time — usually one full cycle of whatever you're testing
• Enough volume — three conversations tell you nothing
• Actual execution — a channel you posted to twice hasn't been tested
• One variable — changing the offer, price and channel together means you learn nothing

If you didn't do the thing properly, you haven't tested it. That's a different finding from "it doesn't work", and it's the more common one.

Distinguish the plan from the strategy

Not every miss means the strategy is wrong. Most don't.

| What missed | Usually means | |---|---| | One quarter's revenue target | Normal variance, or over-optimistic planning | | The same target three quarters running | The plan is wrong | | Leading indicators flat across everything | The strategy may be wrong | | One initiative | That initiative |

A missed quarter is information about your planning, not usually about your business. If you miss the same thing repeatedly, that's a pattern worth acting on — and the useful question is whether the misses share a cause.

When you do change, write down why

Whatever you decide, record it: what you expected, what happened, what you concluded, what you're doing instead.

Two reasons. It stops you re-running the same failed experiment in eighteen months having forgotten the first attempt — which is more common than it sounds. And after several of these you can see whether your predictions are systematically optimistic, which is the same bias correction that makes forecasting useful.

Review these at quarterly planning. A short list of things you tried, what happened, and what you learned is the most valuable planning document a small business accumulates.

The mistakes

1. No falsifier written in advance. Every later judgement is a rationalisation.
2. Watching revenue only. It lags by a full sales cycle.
3. Confusing a wrong idea with wrong execution. Opposite responses.
4. No decision date. You decide when you're tired.
5. Changing course before a fair test. Nothing compounds.
6. Treating one missed quarter as a strategy failure. It's usually planning variance.

What to do next

Take whatever you're currently unsure about and write the falsifier retroactively: what would have to be true by a specific date for this to be working, and what would tell you it isn't.

Then check whether you've already seen that evidence. Frequently you have, and writing it down is what makes it possible to act on.