Five Growth Tests Decision Gate
Fund several bounded experiments, then let results decide whether to scale
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 96%
The method turns a consequential strategy choice into a portfolio of bounded experiments. First define the binary decision, such as whether to raise capital and accelerate growth. Reserve an amount the business can afford to risk, then distribute it across several independent tests aimed at different growth channels or mechanisms. Before launching anything, establish how many tests must succeed to justify the larger commitment. Run the experiments, measure concrete commercial outcomes, and compare the total result with the preset threshold. This prevents either fear or enthusiasm from controlling the decision. Shopify used this mechanism when Tobias Lütke questioned whether the company should remain a profitable lifestyle business or become a venture-backed growth company.
Origin
Tobias Lütke saved $50,000 from Shopify's cash flow and allocated it across five growth tests. He decided that if at least two worked, he would treat that as evidence to raise venture capital. All five succeeded, prompting Shopify to enter full growth mode.
Core principles
- 01Replace conviction debates with bounded experiments
- 02Set the decision threshold before seeing results
- 03Test distinct growth mechanisms rather than minor variations
- 04Treat successful tests as evidence of repeatability
- 05Increase investment only after demand responds
How to run it
- 1
Frame the strategic fork
State the consequential choice in binary terms, such as staying self-funded versus raising growth capital. Define what evidence would reduce uncertainty about that choice.
Pro tip Choose a decision that can genuinely change when new evidence arrives.
Watch out Do not run experiments merely to validate a decision already made.
- 2
Ring-fence the budget
Set aside a fixed amount that the business can lose without threatening operations. Treat it as an evidence-gathering budget rather than general marketing spend.
Pro tip Use a budget large enough to generate a real market response.
Watch out An underpowered test can produce a false negative.
- 3
Build a diversified test portfolio
Design multiple experiments around distinct growth mechanisms, audiences, or channels. Independence makes the combined evidence stronger than repeated versions of one tactic.
Pro tip Include both distribution partnerships and direct customer-acquisition tests when relevant.
Watch out Do not count cosmetic variations of one campaign as separate evidence.
- 4
Set the decision threshold
Decide in advance how many tests must work and what qualifies as success. Record the rule before results can bias interpretation.
Pro tip Tie success to commercial behavior such as activated customers or revenue.
Watch out Changing the threshold after seeing results defeats the mechanism.
- 5
Run and measure
Execute each test within its assigned budget and collect comparable results. Separate promising signals from outcomes that meet the original success definition.
Pro tip Track downstream customer quality, not just sign-ups.
Watch out Vanity metrics can make weak experiments appear successful.
- 6
Follow the evidence
Compare the portfolio with the preset gate and make the strategic decision. If the threshold is met, scale the successful mechanisms and secure the resources required.
Pro tip Preserve the option to revise earlier beliefs without treating that revision as failure.
Watch out Do not scale every experiment simply because the overall gate was passed.
In the wild
Shopify was profitable, had roughly ten employees, and generated more than $1 million in 2008 revenue. Lütke saved $50,000 and ran five growth tests, including a revenue-sharing referral program for web designers and a build-a-business competition. All five worked, giving him evidence that the company could use outside capital productively.
→ Shopify raised a $7 million Series A and shifted into full growth mode.
A self-funded workflow startup allocates £30,000 across partner referrals, outbound sales, a paid implementation pilot, an industry webinar, and a product-led team plan. It decides beforehand that two tests must generate retained accounts at an acceptable acquisition cost. Three pass the gate, so the founders hire a sales lead and raise a focused growth round.
→ The company scales channels supported by observed demand rather than investor enthusiasm alone.
Common mistakes
Testing one channel five ways
Closely related variations do not provide independent evidence that the business has multiple scalable growth paths.
Moving the goalposts
Redefining success after results arrive converts a decision gate into post-hoc storytelling.
Scaling sign-ups instead of value
A test that attracts low-quality users may look successful while weakening retention or unit economics.
Is it for you?
Best for
It is best for cash-generating startups deciding whether to remain small or pursue venture-backed growth.
Not ideal for
It is not ideal when experiments cannot produce meaningful evidence within the available budget or decision window.
From the transcript
“He saves up fifty thousand dollars in cash flow from the company, and he decides I'm gonna use this fifty thousand dollars and I'm gonna…”
“And if maybe like two or more work, then like that's my answer that I should go raise venture capital and pour more fuel on…”
“And of course, all five out of five work.”
From the episode
The Shopify IPO