TThe Foundr Podcast
← All frameworks
InnovationNoah Kagan

Test and Invest

Run a cheap test, then scale only what proves itself

Difficulty
Moderate
Time to result
~weeks to results
Steps
5
Confidence
99%

Test and Invest is AppSumo's resource-allocation rule: ideas can be tried, but they must earn larger investment through a fast, inexpensive test. Begin with a clear hypothesis and success signal, then run the smallest version that can produce useful customer or revenue evidence. If it fails, stop or revise it without exposing the wider business. If it works, invest substantially, with Kagan encouraging teams to ask how they could increase the winner tenfold rather than make a token expansion. He gives video sponsorships as an example: AppSumo sponsored a video, observed strong revenue, repeated the test, and eventually built a five-person team around the channel. He contrasts this with expanding AppSumo's marketplace from 600 to 13,000 products without testing or listening to customers, a decision he says took years to unwind and recover from.

Origin

Kagan presents Test and Invest as an AppSumo operating value and contrasts a successful video-sponsorship test with an untested marketplace expansion. Extracted from The Foundr Podcast.

Core principles

  • 01Earn investment with a small test
  • 02Define success through observed response
  • 03Scale winners materially rather than incrementally
  • 04Keep customer feedback inside the decision loop

How to run it

  1. 1

    State the hypothesis

    Write what you expect the idea to change and for whom. Choose a measurable customer or business response that would support the claim.

    Watch out A test without a success signal can be rationalised after the fact.

  2. 2

    Design the minimum test

    Create the quickest and cheapest version that can produce useful evidence. Limit spend, scope, and exposure while preserving the core proposition.

    Pro tip Use one channel, customer group, or product slice first.

  3. 3

    Read the response

    Measure the chosen signal and gather direct customer feedback. Separate evidence from the team's attachment to the idea.

    Watch out Internal confidence is not a substitute for customer response.

  4. 4

    Stop or invest

    End or revise an unsuccessful test. When the evidence is strong, assign meaningful resources and ask what tenfold expansion would require.

    Pro tip Scale the proven mechanism, not every detail of the pilot.

    Watch out Scaling too early multiplies both cost and customer harm.

  5. 5

    Retest at scale

    Continue measuring economics and customer response as the investment grows. A successful small test is evidence, not permanent immunity from review.

In the wild

Video sponsorship becomes a team

AppSumo sponsored one video and saw strong revenue. The company tested the approach several more times, then built a five-person team to operate video sponsorships.

A limited channel experiment earned a much larger resource commitment.

Marketplace expansion bypasses the test

Kagan expanded AppSumo's marketplace from about 600 products to 13,000 without talking to customers or following the Test and Invest rule. He says customers were angry, the team had opposed the idea, and recovery took substantial time.

The untested expansion damaged the business and had to be reversed.

Common mistakes

Doubling down before testing

Large commitments magnify a weak assumption. Limit the first exposure and require observed evidence.

Scaling by half measures

Once a channel clearly works, a tiny increase may leave most of the opportunity unused. Examine what a material expansion requires.

Ignoring customer objections

Internal enthusiasm cannot override direct customer feedback. Keep customer response in both the test and scale decisions.

Is it for you?

Best for

It is best for marketing, product, and channel decisions that can be tested at limited scale.

Not ideal for

It is not ideal when a tiny test cannot represent the real outcome or when failure would still be irreversible.

From the transcript

people can pretty much do 31 00 whatever they want but they have to test it first

Noah Kagan · (31:00)

If it works, how do we double down on it, 10x it or whatever that is?

Noah Kagan · (31:00)

From the episode

498: He Lost Millions in Facebook Stock Then Built a $300M Tech Company

Noah Kagan