Welcome to Language Agnostic, the blog of Inaimathi! It's now built on top of the Clojure framework http-kit. I'm reasonably confident it's no longer the least performant blog on the internet.

Enjoy the various programming-themed writings availble on offer. The latest post is available below and the archive link is directly above this text.

Antler - Day 8

Sat Apr 13, 2024

I wasn't here day 7. It happens, but hopefully not too frequently.

Keynote - Rio Hodges

  • History consisted of a few startups, and moved into investment
  • Has talked to customers about lots of things. Health concerns, pets, zoning, all of it. And he helps Antler US with customer discovery.
  • How do I do customer discovery?
  • What do I do afterwards?
  • How do I avoid passing the Mom test?

Customer Discovery

The fastest way to build a startup is to ignore everyone and go non-verbal, work on your computer at home. The best way to build one is to constantly look for feedback in the earliest stages and iterate before over-building.

There's three levels of customer discovery:

  • pay for a survey, or build your own to get initial data, and do some market research
  • Problems
    1. People don't take surveys that seriously
    2. Customers rarely know exactly what they want and don't want if prompted
    3. Great for generating volume, but not a dense quality of information
  • Run 1:1 customer discovery calls. 30-minutes max, paid or unpaid. Build a concise discussion guide before hand. Or, run focus groups. No more than 6 people, facilitate a 1-hour discussion.
  • Problems:
    1. You'll always pass The Mom Test (people always want to say your idea is great)
    2. Mediating groups isn't easy without prior experience
    3. Fear of your idea getting shot down. It's hard not to have this, but ideally, you want to completely avoid the sunk-cost fallacy.
  • Sit and watch someone do what they would normally do. Have them zoom plus screen-share, do an in-person demo or sitting over their literal shoulder. Use a duct-taped demo and have them do something.
  • Problems:
    1. As a customer, it's difficult to not get stage fright
    2. Takes a lot of time and energy
    3. Using something duct taped can force users into UI/UX issues that don't contribute to the high level idea

Realistically, you'll need to do all three of these at some point (the Hard mode sounds like something you definitely want when you get to UI development). Your role as a founder is to constantly talk to customers and iterate. This is why tenacity is really important as a customer.

What do you do next?

  • Build and test product to meet customer needs
  • Work towards discovering Insights
  • An Insight is not:
    • a fact
    • a conjecture
    • obvious
  • "An Insight is the capacity to extract an ought from an is." (I don't think he meant philosophically)
  • Example of not Insight:
    • "People want food delivery to be fast"
    • "People don't like working in cubicles"
    • "People often buy exercise gear that they don't use"
  • An Insight is a realization that helps us see a business or product in a new light. Used to inform hypotheses, features, strategies.
  • An Insight moves you from a "Could" to a "Should" unambiguously in the space of idea evaluation
  • There are three ategories of Insights:
    • Company level
    • Industry level
    • Channel level

Examples of Insight

  • Delta: commercial flying is a race to the bottom (industry level insight). Rewarding loyalty generously creates a consumer that will trust you across their spend (company level insight). Their loyalty program/amenities is probably more profitable than their flights at this ponit.
  • Peloton: Fitness companies don't compete with other gyms, they compete with inaction that prevents people from going to the gym in the first place.
  • AirBnB: A market is only as strong as its' weakest interaction and your only recourse is to solve a problem after it happens.

So the insight isn't "people want an easy way to rent out their homes" or "people want an internet-connected bike" or "people want cool lounges". Those follow from the insight, which connects a tendency of the market to a mechanism you could construct to make it more efficient.

Example process

Start with a fact: "No one likes doing expenses"

  • This is a fact, it doesn't convert a Could to a Should, it just points to an Is.

Accepted belief: Expenses are one of those chores you'll eventually have to get around to.

  • This is a current tendency of the market. There's a thing a lot of people have do, and are annoyed by, that ends up dictating their behavior in a suboptimal-on-reflectoin way.

Insight: Expenses never look or feel convenient, so we do the only thing we can to alleviate our pain. We procrastinate.

  • The company Ramp realized that their position was basically an anti-akrasia tool. Their angle of attack was not to put together a web platform, they basically allow you to use their system however you need to in order to settle your expenses (SMS/mobile app/call/wherever you're at), and they push hard to get you over the hump of procrastination.

How do you avoid passing The Mom Test

  • The Mom Test: The incredibly easy trap of being told - as if you were talking to your own mother - how wonderful, useful and incredible your startup idea is.
  • Examples of terrible questions:
    • Do you like this product?
    • Would you like the product if it did <insert features>?
    • How much would you pay for this product?
    • It's like the Uber for X, does that make sense to you? _(Metaphors are imperfect, and this one plays on exisitng, intrin)
    • I'm building something that does <thing company does>, do you think that will work for you? (Beginning of a soft sell)
  • Instead of pitching your startup; you need to deeply understand your customer.
  • Seek to
    1. Talk about your customers' life
    2. Listen, don't pitch
    3. Get specifics, not generics
  • Examples of better questions:
    • Talk me through the last time you went through this problem... in detail?
    • What else have you tried? Why didn't it work?
    • Why haven't you solved this problem yourself?
    • How does this fit into your existing workflow?


  • Regarding the Mom Test; what's your perspective on going to meetups or other events like the author suggests in the book?
    • Depending on the type of meetup, it's a great place to make connections in a space. The meetup itself shouldn't be a call for you to sell your startup. Both because it's annoying in meetup spaces, and because it basically guarantees that your best-case scenario is passing the Mom Test.
  • How far into the process should you tell them your solution? You don't want to mislead them, so when should you tell them what you're doing?
    • So, probably in the fourth quarter, you should tell them. And ideally, don't ask any of the terrible questions (even implicitly), but rather give them a demo. Either an API key or something or a video. I've seen people successfully turn their customer discovery into a sales opportunity with something like "Hey, I'm tinkering with this thing , it's not ready yet but would you be interested in getting updates?" But that should be it. Basically anything stronger than this runs you up against the Mom Test; it's surprisingly common in any context where you're tempted to show initiative or advocate for your idea.
  • If we're doing cold outreach, what are strategies to get a higher conversion rate?
    • If the problem is deep enough, they'll want it solved in some way. I've seen people incentivize customer discovery calls with gift cards (or like, "Hey, I'll buy you a coffee"). Be direct about it; "Hey, I'm an entrepreneur working in <this space>, I see you're <someone associated with the space>, can I get you coffee and chat you up for half an hour?". Also, warm connections are better because there's social capital involved, and hence higher stakes if you fuck something up. Check in with your friends/Antler colleagues.

Exercise - Build your own discussion guide

  • What does a good discussion guide look like?
  • (2 minutes) Identify your ICP (initial customer profile)
    • occupation, role
  • (10 minutes) Create a minimum of 7 questions you would like to ask your customer
    • Make sure they're not leading, asking for validation, or passing The Mom Test
  • (20 minutes) Share with a neighbor and provide feedback.
    • Is the question specific enough?
    • Are you leading the witness?

Creative Commons License

all articles at langnostic are licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License

Reprint, rehost and distribute freely (even for profit), but attribute the work and allow your readers the same freedoms. Here's a license widget you can use.

The menu background image is Jewel Wash, taken from Dan Zen's flickr stream and released under a CC-BY license