CaddieOpen in CaddieDownload for Mac

How to test for real customer pull early

Shared by Caddie Library · Updated
Chapters show their main points
Chapter 1 of 15 · Finding Product-Market Fit
← The growth course
1. Finding Product-Market Fit
How to test for real customer pull earlyHow to pick a narrow niche to start inHow to launch a simple first version fastHow to learn from users and customersHow to read early signs of product-market fit
2. Onboarding and Activation
How to structure onboarding as a storyHow to ask and use onboarding questionsHow to get users to the aha moment fastHow to decide how long onboarding should beHow to build quiz funnels for paid traffic
3. Paywalls and Pricing
When to choose a hard paywall over freemiumHow to design a paywall that convertsHow to price annual and monthly plansHow to set, raise and localize pricesHow to run free trials and free tastes
4. Business Models and Revenue
How to structure subscription tiersHow to keep a free tier that pays offHow to charge for usage and AI creditsHow to pair hardware with a subscriptionHow to add ads, affiliates and partners
5. Retention and Engagement
How to measure retention honestlyHow to build habit loops that bring users backHow to use notifications and lifecycle messagesHow to reduce churn and handle cancellationsHow to build social loops and community
6. Brand and Positioning
How to position against bigger incumbentsHow to name and describe your productHow to earn trust in high-stakes categoriesHow to borrow credibility from celebrities and experts
7. Organic Content and Creators
How to find, vet and recruit creatorsHow to pay creators fairly and safelyHow to brief and manage creatorsHow to find and copy viral formatsHow to scale one format across many accountsHow to rank in AI search and SEO
8. Paid, App Store and Viral Growth
How to make ads that convertHow to turn organic winners into paid adsHow to structure and scale Meta campaignsHow to tell if paid acquisition is profitableHow to win in the App StoreHow to build viral and referral loops
9. Experimentation and Analytics
How to run trustworthy experimentsHow to judge test results correctlyHow to run more experiments fasterWhat metrics to track and how to define them
10. Building AI Products
How to build for the models of tomorrowHow to design AI interfaces beyond chatHow to design AI agents people can trustHow to measure AI quality with evalsHow to control AI costs and pricingHow to build products with AI coding tools
11. Teams, Hiring and Funding
How to hire people who thrive on ambiguityHow to organize small product teamsHow to run exploratory bets and kill themHow to work as a team with AI agentsWhen to raise money or sell your company
12. Founder Principles
How to pick a mission worth building a company aroundWhy to trust your own judgment over the consensusHow to turn failure and setbacks into the next stepHow to learn from rivals and copy what works
13. Choosing Where to Compete
How to make low cost your competitive advantageHow to create a market instead of fighting for shareWhy scarcity and brand can beat volumeHow to build an advantage that feeds itselfHow to attack a bigger rival where it cannot followWhen to stay focused on the core business
14. Building and Leading the Company
How to hire great people and give them roomHow to align pay and ownership with the missionHow to keep your culture and standards as you growHow to build candor and fast decision-makingWhy leaders must stay close to the work and customersHow to choose and prepare a successor
15. Capital and Deals
How to fund growth without losing controlHow to acquire companies and keep what makes them workWhat to do with cash: reinvest, buy back or holdWhy to keep a cash cushion and avoid heavy debtHow to structure ownership to protect your missionHow to negotiate deals that keep your leverage

1 The bottom line

Test pull by asking a specific person to pay, or commit real effort, for a manual or fake version before you build. Stated interest is cheap, and payment shows who truly needs it.

2 Key takeaways

  • Write your pull hypothesis for one named person with a project, a deadline and a gap in today's options. A buyer needs all four PULL parts, and missing one gives you complaints or lukewarm interest.
    • Jump's AI note taker had all four: notes are legally required, and generic note takers lacked CRM integration and SEC compliance. Insurance agents had no legal requirement to log notes, so the pull was missing Rob Snyder13:30.
  • Ask for money before you build, even a small amount. A payment is a stronger signal than any answer in an interview.
  • Deliver the promise by hand to a few buyers first. A manual pilot tests the real experience with no code.
    • BoldVoice recruited a speech coach and about a dozen beta users and ran it by email. The pilot led to a first term sheet of around $500K and a YC acceptance Anada Lakra, BoldVoice11:40.
  • Discount polite interest and friends' feedback, and count only buying signals. Friends and kind prospects say they like it even when they will never buy.
  • Plan for rejection and keep your call volume high. Most customers are not early adopters and have no reason to take risks.

the true signal of product market fit is ultimately having people open up their wallets and pay you

Naomi Ionita, Evernote

2.1 Write a pull hypothesis for one person

Pull needs all four: Project, Unavoidable, List, Limitations
Write the hypothesis for one named person with one project. If any slot stays empty, expect polite interest instead of a purchase.

A pull hypothesis names one buyer, their project, and why today's options fail them.

  • Fill in the four PULL parts as a Mad Libs sentence. Rob Snyder13:30
    • Write down the Project on their to-do list, the Unavoidable timing, the List of options they considered, and the Limitations that make those options not good enough Rob Snyder13:30.
    • Write it for a specific person, then use the calls to check each part Rob Snyder13:30.
    • If one part is missing, expect the buyer to complain, stay lukewarm, or say they will do it themselves Rob Snyder13:30.
  • Name the role and the task, not the company. Rob Snyder16:35
    • Pull belongs to a person with a project they cannot do themselves, not to a segment Rob Snyder16:35.
    • "You're not selling to SpaceX, you're selling to the head of legal at SpaceX." Another example is an operating partner at a mid-market private equity firm trying to do X, Y and Z Rob Snyder16:35.
  • Look for signs of the problem where people already ask for help. Rafael Kramer17:47
  • Score your product against the old way before you commit. Kunal Shah, CRED5:09
    • Ask users to rate the existing solution out of 10, then rate yours Kunal Shah, CRED5:09.
    • At a gap of 4 or more, the switch is irreversible, users tolerate occasional failure, and they brag about it. Below 4, switching stays reversible Kunal Shah, CRED5:09.
    • Uber scores about 9 against a cab at 3. Buying a suit online scored 5 against 7 offline Kunal Shah, CRED5:09.

2.2 Run short sales conversations

Many short calls with the right person show fast whether the idea pulls.

  • Book calls at high volume. Raaz Herzberg, Wiz12:32
    • Wiz held 10 to 15 calls a day with CISOs, presenting the idea and listening for reactions. Volume is how you separate an idea that isn't working from one that is Raaz Herzberg, Wiz12:32.
    • Audit your calendar for the past month and count physical meetings with potential customers. Aim for 20 to 30 percent of your time Dalton Caldwell, Y Combinator41:21.
    • The calls showed the original network security idea drew only polite interest. That led to a pivot six weeks after Raaz joined Raaz Herzberg, Wiz12:32.
    • Zip's founders spent well over 20% of their time on customer calls, did hundreds of them, and pre-sold before building Dalton Caldwell, Y Combinator41:21.
  • Get the calls with a short cold note about the problem. Jen Abel, Jellyfish12:53
    • Open with why it matters to their role, then say something counterintuitive. Talk about the problem and why today's options fall short, not your solution Jen Abel, Jellyfish12:53.
    • Keep it to three or four sentences so it fits one phone screen Jen Abel, Jellyfish13:39.
    • Jen's first 20 Jellyfish customers came from cold email using the line "zero to one sales talent doesn't exist" Jen Abel, Jellyfish12:53.
  • Pitch your real sales story, then ask if they would pay and why. Madhavan Ramanujam, Simon-Kucher13:00
    • Use the same story you would use after launch, and repeat the exercise as the product evolves. There is no such thing as too early Madhavan Ramanujam, Simon-Kucher13:00.
    • Try the narrative in a few calls before the whole team uses it. Ask "am I crazy or are you seeing this?" and watch whether prospects nod and volunteer how the shift plays out for them Andy Raskin50:10.
    • Simon-Kucher benchmarked that 72 percent of innovations fail commercially, largely because pricing was an afterthought Madhavan Ramanujam, Simon-Kucher13:00.
  • Follow the customer who is surprisingly easy to sell to. Brendan Foody, Mercor43:26
    • If the marginal customer is very hard to sell, you can't build a huge business. Look for a large pain point and an easy sale Brendan Foody, Mercor43:26.
    • Mercor's pull came from an xAI founders meeting and a crowdsourcing player hiring over a thousand people on its platform Brendan Foody, Mercor43:26.

2.3 Pre-sell with a real payment

A payment, even a small or refundable one, shows pull better than any answer.

2.4 Run a paid manual pilot

Deliver the promise by hand to a few buyers and see whether they pay and come back.

2.5 Read the signals honestly

Count payments and sharp buying questions, and discount politeness and friendly feedback.

2.6 Where they disagree

Advice splits on free trials, on pages versus live calls, and on how much inbound interest proves.

  • Should the paid test include a free trial?
  • Is a page enough, or do you need live conversations?
  • Do inbound requests and interview enthusiasm prove pull?
    • Grammarly shipped what users asked for, expecting the tip of the iceberg. The floodgates did not open Yuriy Timen, Grammarly28:56.
    • RevenueCat listed nine integrations it had not built, and built each one over a weekend only when someone asked Jacob Eiting, RevenueCat9:40.
    • In prototype interviews, Chris acted as if the product was live. Users who were disappointed it was not out signalled pull. Yet Autopilot did not become an organic growth channel, because people don't automate their financial lives today Chris Hutchins, All the Hacks10:30.
Next lesson →How to pick a narrow niche to start in

Sources

Part of Finding Product-Market Fit in the growth library.