Why Do I Keep Changing My Business Idea?

Why Do I Keep Changing My Business Idea?

Short answer: You usually keep changing ideas because you compare imagined upside instead of comparable market evidence, or because switching lets you avoid asking anyone to pay. 

Choose one idea for a 14-day test, define the result that would make you continue or stop, and reject new ideas until that test ends. Change direction when evidence fails a pre-set threshold, not when another idea feels more exciting.

Changing an idea can be intelligent. Customers may reveal that the problem is weak, the buyer is unreachable or the price cannot support delivery. The damaging pattern is changing before the idea has faced the same standard of evidence as its replacement.

A new idea arrives without objections, failed conversations or awkward arithmetic. It therefore looks cleaner than the idea you have started testing. You are comparing a tested weakness with an untested fantasy.

Use the Idea-Switching Diagnostic Tree

The Idea-Switching Diagnostic Tree identifies the reason for a proposed change. Start at the top each time you feel pulled towards another idea.

The tree makes changing expensive in evidence rather than shame. You do not need to feel certain. You need a recorded reason that another competent person could inspect.

Stop choosing identities and choose a test

“I could run a design studio” invites you to imagine a future identity. “I will ask 20 independent cafés whether they will pay £250 for a fixed menu redesign delivered in five days” creates a decision.

An idea becomes testable when it contains four parts:

1. One identifiable buyer.
2. A problem the buyer already recognises.
3. A result with a defined boundary and price.
4. An action that exposes the offer to a buying decision.

If you cannot state those parts, the problem is not that you have too many ideas. None of them is yet specific enough to compare.

My view is that you should choose the idea with the cheapest honest route to rejection. Founders often choose the idea with the largest theoretical market. Early on, fast disproof is more valuable because it protects months of cash and effort.

Give every idea the same evidence allowance

One idea may have received 30 cold messages and three difficult calls. Another may have received praise from two friends. Treating the second as more promising would reward weaker evidence.

Create a common test allowance. For a simple service, that might be 20 suitable prospects, at least five conversations and a direct request for a paid trial. For a physical product, it might be ten conversations followed by a specified number of paid orders or deposits before production. The right threshold depends on price, market access and buying cycle, but it must be set before results arrive.

Do not change the offer, buyer and channel simultaneously halfway through. You will not know which assumption produced the result. Make one deliberate revision, record it and begin a fresh round.

Separate a pivot from an escape

A pivot preserves evidence. If restaurant owners reject your stock-counting service but repeatedly pay for waste analysis, moving to waste analysis follows the customer problem. If you abandon restaurants for handmade jewellery because outreach felt uncomfortable, you have escaped the test.

Use this distinction:

A new idea is allowed. It simply returns you to the beginning. Do not claim that progress on the previous idea reduces the new idea's risk unless the evidence genuinely transfers.

Worked example: Danielle's inventory photography service

Danielle cycles through podcast editing, mobile car cleaning and personalised prints. She spends ten hours researching each. At an opportunity value of £20 an hour, 30 hours of switching has cost her £600, although no cash left her account.

She commits to one offer: inventory photography for small letting agents, priced as a £120 paid trial. During 14 days, she contacts 20 relevant agents, has four conversations and wins two trials. Each trial costs £22 in travel and takes three hours to shoot and deliver. Outreach and calls took eight hours in total.

Her test arithmetic is:

• Revenue: 2 × £120 = £240
• Direct cost: 2 × £22 = £44
• Cash contribution: £240 minus £44 = £196
• Total time: 8 outreach hours + 6 delivery hours = 14 hours
• Effective return: £196 divided by 14 = £14 an hour

Danielle requires £25 an hour, so the current offer fails on pay. That does not justify an unrelated fourth idea.

She has evidence that some agents buy, plus evidence that £120 is too low for her present delivery time.

She can run one controlled revision at £180. At two sales, cash contribution would be £360 minus £44 = £316, or £22.57 an hour across the same 14 hours. It still misses £25, so she would also need faster delivery, lower travel or a higher price. The arithmetic tells her what to change and when to stop.

Create an idea parking rule

Write new ideas in one line, with the buyer and problem if known, then leave them untouched until the current test date. Do not research names, suppliers or branding. Those actions create the feeling of progress while consuming the attention needed for a buying test.

At the review date, score the current and parked ideas against the same criteria: problem evidence, buyer access, cash before proof, contribution potential and fit with your available time. 

A parked idea can win, but only after the current evidence is recorded.

This rule preserves creativity without giving every thought the right to interrupt operations.

Finish one 14-day decision cycle

Today, choose the idea that can reach a genuine buying decision with the least irreversible spending. Write its buyer, problem, result and price on one page. Set one continuation threshold and one stopping threshold.

Over the next 14 days:

1. Approach the defined number of suitable buyers through one primary route.
2. Record responses, conversations, objections, payments, costs and hours.
3. Make no more than one controlled change to the offer.
4. Park every new idea without researching it.
5. On day 14, continue, revise once or stop according to the thresholds you set.

If you cannot complete the uncomfortable sales action, name avoidance as the result. Do not blame the idea for a test it never received.

Frequently asked questions

How long should I stick with one business idea?

Stick with it until you complete a pre-defined evidence cycle, not for an arbitrary number of months. A straightforward local service might produce interpretable evidence after 20 suitable approaches, five conversations and direct paid offers. A high-value B2B sale may require longer because the buying cycle is slower. 

Set the volume, timeframe and decision threshold before starting. Continue when buyers take the next commercial step and the economics can work. Stop or revise when the same material objection repeats despite reaching the right buyers. The exception is new legal, safety or cost information, which can justify stopping immediately.

Is it sensible to test several ideas at once?

Usually not if you have limited time. Parallel tests divide prospecting effort and make weak response difficult to interpret. Run one primary test until it reaches its evidence allowance. You can compare two ideas only when the tests use similar effort, buyer access and commitment signals, and when running both does not delay either.

Record hours as well as cash. Ten casual messages for each of four ideas are not equivalent to a serious test. The exception is a tightly controlled comparison, such as offering two distinct results to the same buyer group, provided you do not confuse buyer preference with unclear messaging.

What if I have already spent money on the current idea?

Ignore money that cannot be recovered when choosing the next test. List what remains usable, saleable or refundable, then compare the future cash and effort required by each option. A £1,000 website already paid for does not make another £2,000 campaign sensible. 

Equipment with a credible resale value may reduce the cost of stopping, while customer deposits or contracts may create obligations you must still fulfil. Record the lesson from the spending, but do not make new decisions merely to justify it. The exception is a cost attached to a legal, safety or customer duty, which needs qualified local advice before you withdraw.

Does losing interest mean the idea is wrong for me?

Not necessarily. Interest commonly falls when work moves from imagining to prospecting, because rejection and repetition replace novelty. Identify the exact task you resist. If you dislike selling, every business will expose some version of that problem. If you dislike the core delivery after several representative jobs, the fit may genuinely be poor. 

Compare your energy after customer work, administration and learning separately. Do not demand constant excitement. The exception is a model that conflicts with your values, health or essential commitments. Those are valid constraints and should be included in your written decision criteria rather than treated as a failure of discipline.

Can I return to an idea I rejected before?

Yes, when a specific assumption has changed and you can name the new evidence. Record why the earlier test stopped, such as an unworkable price, inaccessible buyer or excessive delivery time. Then show what is different now: a lower supplier minimum, a warm route to the buyer, a narrower scope or a verified change in demand. 

Do not revive the idea because you remember its upside more clearly than its failure. Set a fresh threshold and retest the changed assumption first. If nothing material has changed, returning is another form of switching, not a new opportunity.

What if my business partner and I prefer different ideas?

Replace preference with a common test and agree the decision rule in writing. Each person should state the buyer, problem, price, cash required before proof, delivery hours and evidence target for their option. If feasible, run equal tests sequentially or in parallel with the same time allowance.

Decide in advance what happens if one wins demand but loses on economics. A disagreement may also reveal incompatible goals, such as one partner wanting quick income while the other wants a long product build. Resolve that objective before choosing an idea. For ownership, duties or exit decisions, use qualified local legal advice.