What founders, investors and authors say works — summarised in plain words, linked to
the source, and tried on the businesses we run. When enough of them have tried it, we say whether it worked.
121 classics · all free0 being tested0 with a verdict5 businesses before a verdict
Paul Graham: Co-founded Y Combinator, the startup program behind Airbnb, Dropbox, Stripe and Reddit.
“Make something people want.”
A new company's one essential job is to build something a real group of people actually wants; if that works, making money from it is the easier problem and can come after. Graham pairs the motto with advice not to worry much about the business model at first.
When it applies: You are choosing what to build, or deciding whether the current offer deserves more effort.
Not tried yet
None of our businesses has tried it yet.
Startup Playbook: a few users who love it beat many who like it
Sam Altman: CEO of OpenAI, the maker of ChatGPT; formerly president of Y Combinator.
“It’s much better to first make a product a small number of users love than a product that a large number of users like.”
Altman's summary of YC advice: first build something a small group loves, because finding more users is easier than turning 'like' into 'love'. Run a tight loop of talking to users, watching them use the product and fixing what is weak; keep founders doing sales and support; launch something simple sooner than feels comfortable.
When it applies: You have early users and must decide whether to widen reach or deepen the product for the users you already have.
Not tried yet
None of our businesses has tried it yet.
Launch something bad, quickly (plan an MVP in weeks)
Michael Seibel: Co-founded Justin.tv, which became Twitch; led Y Combinator's startup program from 2016 to 2024.
“Hold the problem you're solving tightly, hold the customer tightly, hold the solution you're building loosely.”
A pre-launch company's goal is to get a very simple version in front of real users fast, then get their feedback and iterate. Time-box the first spec to what can be built in a few weeks, write it down so it does not drift, talk to a few users before building, and do not ask users to design features for you.
When it applies: You are still building before launch, or the first version keeps growing in scope.
Not tried yet
None of our businesses has tried it yet.
If you're not embarrassed by v1, you launched too late
Reid Hoffman: Co-founded LinkedIn; early PayPal executive and early investor in Facebook and Airbnb.
“If you're not embarrassed by the first version of your product, you've launched too late”
Launch fast, because some of your assumptions about customers will be wrong and learning only starts once real users arrive. 'Embarrassed' means mildly so: a launch that causes lawsuits, drives users away or burns money for nothing was too early.
When it applies: You are holding back a launch to polish features whose value users have not yet confirmed.
Contested. In some markets a bad first impression sticks (anything touching trust, money, health or security). Hoffman himself rules out launches that harm users, invite lawsuits or waste resources.
Not tried yet
None of our businesses has tried it yet.
Rahul Vohra: Co-founded the Superhuman email app; earlier co-founded Rapportive, which LinkedIn bought.
“spend half your time doubling down on what users already love and the other half on addressing what’s holding others back”
Turn the Ellis survey into a repeating loop: survey users, narrow your target to the kind of people who answered 'very disappointed', learn what they love and what holds the 'somewhat disappointed' back, split the roadmap between those two, set aside the 'not disappointed', and track the score as the main metric. Superhuman reports rising from 22% to 58% this way.
When it applies: You have some engaged users (Vohra says about 40 responses give directional results) but fewer than 40% would be very disappointed to lose the product.
Not tried yet
None of our businesses has tried it yet.
Jobs to be done: customers hire products to make progress
Clayton M. Christensen: Harvard professor whose theory of 'disruptive innovation' changed how companies see upstart rivals.
People buy not because of who they are but because they are trying to make progress in a particular situation: they 'hire' a product for that job and 'fire' it if it does the job badly. Jobs have social and emotional sides as well as practical ones, so design the offer and the whole experience around the job rather than around customer demographics.
When it applies: Sales are weak despite a decent product, or you are deciding what to build, how to describe it and whom to target.
Not tried yet
None of our businesses has tried it yet.
Joel Spolsky: Co-founded Stack Overflow; his company Fog Creek created Trello.
“Smart companies try to commoditize their products’ complements.”
When another product is used together with yours, demand for yours rises as that other product gets cheaper. So companies push the price of complements toward zero, by giving them away, open-sourcing them or encouraging many cheap suppliers, in order to sell more of the thing they actually profit from.
When it applies: Customers need a second product or service alongside yours, and its cost or hassle is holding back purchases of your main offer.
Not tried yet
None of our businesses has tried it yet.
Seth Godin: Bestselling marketing author; sold his early internet company, Yoyodyne, to Yahoo.
Interruptive advertising keeps losing power, so the marketing has to be built into the product: make something unusual enough that people notice it and tell others. A safe, average product stays invisible however much is spent promoting it.
When it applies: A product enters a crowded category where buyers already have adequate choices and tune out ads.
Not tried yet
None of our businesses has tried it yet.
The three kinds of platforms: access API, plug-in, runtime
Marc Andreessen: Co-created Mosaic, the first popular web browser, and Netscape; co-founded Andreessen Horowitz.
“If you can program it, then it’s a platform. If you can’t, then it’s not.”
Only something outside developers can program is a platform, and it comes in three levels: an access API that outside apps call, a plug-in API that lets outside apps appear inside your product, and a runtime that hosts outside developers' code for them. Each level is harder for the owner to build but more useful to developers than the last.
When it applies: Deciding whether, and how deeply, to open a product to outside developers.
Can't test here
It needs an outside developer community building on the product, which a small business cannot attract and measure within months.
Steve Jobs: Co-founded Apple; led the Mac, iPod and iPhone, and ran Pixar.
“Focusing is about saying no.”
A company does better by cutting most of its projects and features and putting its effort into a few, even when the ideas it drops are good ones. Jobs argued that Apple had spread itself across so many directions that the whole added up to less than the parts.
When it applies: A business is running several products, offers or features at once and none of them is clearly winning.
Not tried yet
None of our businesses has tried it yet.
Start with the customer experience and work back to the technology
Steve Jobs: Co-founded Apple; led the Mac, iPod and iPhone, and ran Pixar.
“you've got to start with the customer experience and work backwards to the technology.”
First decide what the customer should experience and get out of the product, then choose or build the technology that delivers it. If you start from a technology and then look for someone to sell it to, you tend to build things nobody wants. Jobs admitted he had made that mistake himself.
When it applies: Before building a new product or feature, especially when the excitement is about a capability (a new model, an integration) and not about a customer problem.
Not tried yet
None of our businesses has tried it yet.
Steve Jobs: Co-founded Apple; led the Mac, iPod and iPhone, and ran Pixar.
“Real Artists Ship”
Work only counts once customers have it. Keep the launch date and finish what you can, rather than letting it slip for more polish. In January 1984 Jobs turned down the Mac software team's request for a two-week delay, and the team shipped on time.
When it applies: A launch keeps slipping because the product never feels finished.
Not tried yet
None of our businesses has tried it yet.
Elon Musk: Runs Tesla and SpaceX; co-founded the company that became PayPal.
“boil things down to their fundamental truths and reason up from there, as opposed to reasoning by analogy.”
Most decisions copy what others already do, with small changes, and that is fine for everyday choices. For something new, break the problem down to facts you are sure of (what the parts really cost, what the customer really needs) and build the answer up from those, without assuming the usual way is right.
When it applies: A business is about to copy a competitor's setup, cost structure or offer only because that is how the category usually works.
Not tried yet
None of our businesses has tried it yet.
Working backwards: write the press release and FAQ before you build
Colin Bryar: Amazon executive for twelve years, including two as Jeff Bezos's chief of staff.
Bill Carr: Amazon executive who launched its digital music and video businesses.
Before committing to build something, write a short mock press release announcing it to customers, plus a list of the hard questions and their answers. If the announcement would not excite a customer, or the answers expose weak spots, change or drop the idea before money is spent.
When it applies: Choosing which new product, feature or service to build next.
Not tried yet
None of our businesses has tried it yet.
Shape Up: fixed six-week cycles set by appetite, not estimates
Ryan Singer: Spent 17 years at Basecamp; wrote Shape Up as its head of product strategy.
“Estimates start with a design and end with a number. Appetites start with a number and end with a design.”
Decide up front how much time an idea is worth (its 'appetite') and shape the work to fit. Then give a small team a fixed cycle (six weeks at Basecamp) to finish it, cutting scope instead of moving the deadline. Unfinished work is not extended by default (the 'circuit breaker'), and there is no standing backlog: ideas are pitched again each cycle.
When it applies: Projects routinely overrun their estimates or never quite finish.
Not tried yet
None of our businesses has tried it yet.
We don't sell saddles here: sell the outcome, not the features
Stewart Butterfield: Co-founded Slack and the photo-sharing site Flickr.
“That’s why what we’re selling is organizational transformation.”
Few people shop for a category of software; they want the result it brings. So describe and sell the change in the customer's life (less email, calmer teams), not the feature list. Look at the product the way a first-time user would, and remove every rough edge.
When it applies: A business sells something new, or competes in a crowded category, where buyers are not searching for the product by name.
Not tried yet
None of our businesses has tried it yet.