← Back to the principles list

Planning Fallacy in ecommerce

Next principle →

Trusting the plan alone often leads to jobs that run long and cost more.

A rebuild planned for eight weeks sounds manageable, so you say yes. Ask how long similar rebuilds really took, and you may see a very different number before you've paid anything.

Planning fallacy
Made-up example: showing the usual time changes which redesign owners choose.

The planning fallacy, a form of optimism bias, is the tendency to guess too little time and money for a future job, even when similar jobs have run over before.

Before you agree to a redesign or launch, look up what similar jobs really took and plan from that. When someone quotes you a date, ask for the finish dates on their last few jobs. Then compare your options on the usual time as well as the promised time.

Planning fallacy example: an 8-week rebuild that takes 7 months

A store owner is choosing a redesign. One option is a custom rebuild with an 8-week plan at $15,000. The other is a theme redesign, done in 72 hours, at $99. Going by the plan alone, most owners pick the rebuild, which takes 7 months and costs $43,000. Now add one fact to the rebuild's line: “8-week plan, usually 7 months.” Everything else stays the same, and most owners pick the theme redesign, which takes 72 hours and costs $99. This is a made-up example, and a theme redesign won't suit every store. Ask how long a job usually takes, and trust that over the plan.

How the planning fallacy works with other principles: 4 store examples

A late project can have more than one cause. One of these pairs is a common mix-up, two make a bad estimate worse and one keeps a late job going. Every store below is made up to show the idea.

Planning fallacy vs Parkinson’s law

Often confused with: The planning fallacy is guessing too little time at the start, while Parkinson’s law has work growing into whatever time you allow.

Illustrative store example: A board game store owner guesses that new how-to-play pages will take one week, and they take four, which is the planning fallacy. Given a full month for the next batch, the team keeps polishing until the month is gone, and that is Parkinson’s law.

Parkinson’s law: why work grows to fill the time allowed →

Planning fallacy and the Dunning-Kruger effect

Works with: A merchant new to Liquid, Shopify's theme code, can overrate their skill, which is the debated Dunning-Kruger pattern, and so guess too little time for the work.

Illustrative store example: A surf shop owner watches a few theme coding videos and plans to rebuild the product page over a weekend. Three weekends later the phone layout is still broken, so ask someone experienced to look at the plan before you set a date.

Dunning-Kruger effect: why beginners can misjudge hard work →

Planning fallacy and hindsight bias

Works with: Looking back, the delays on your last project can seem obvious and avoidable, so you plan the next one as if they won't happen.

Illustrative store example: A flower shop's last site update ran five weeks over because the photos came late, and the owner now says anyone could have seen that coming. She plans the next update as if late photos were the only risk, when the safer starting point is the full time the last update took.

Hindsight bias: why past problems look obvious afterwards →

Planning fallacy and the sunk cost effect

Works with: The planning fallacy can leave a job over budget, and then the money already spent can push the merchant to keep paying.

Illustrative store example: A furniture store budgeted $8,000 for a new site and has spent $20,000 with no finish date in sight. Ignore what's gone and ask whether the remaining cost and time are still worth it compared with starting something smaller.

Sunk cost effect: why money already spent keeps you going →

Put it into practice

How do I make a more honest estimate for a store project?

Start from past jobs and set your plan aside for a moment. List two or three similar projects, yours or other owners', and how long each took from start to finish. Use that as your estimate, then adjust only for what's truly different this time. Splitting a big job helps too: finish one page first, and you'll have a real number for the rest.

Can a longer estimate cause its own problems?

Yes. Add too much spare time and the work can spread out to fill it, so set checkpoints along the way. A long estimate can also scare you off a job worth doing. A custom rebuild is still the right call for some stores, as long as you choose it knowing its real length and price.

How is the planning fallacy different from Parkinson’s law and scope creep?

The planning fallacy is guessing too little at the start. Parkinson’s law is work stretching to fill whatever time you allow. Scope creep is the job itself growing after it begins, as new pages and requests get added. One project can suffer all three, so write down what's included before you begin and check that list when the job slips.

How do I know my estimates are improving?

Keep a simple list of each job, your guess and what it really took in time and money. After a few jobs you'll see how far off you usually are. If your jobs tend to take twice your guess, use that to correct the next one. You're improving when the gap between the guess and the real number keeps shrinking.

Why does a draft theme help when a redesign runs late?

A draft theme is an unpublished copy of your store's design, so work on it doesn't touch the store your shoppers are using. If the job runs late, your live theme keeps running as it is. That takes away the pressure to publish half-finished pages. It won't shorten the work or fix a bad estimate, so you still need a realistic date.