Frameworks for startups

Max Heinritz
7 min readJun 10, 2018

Below are the lenses through which I evaluate professional opportunities. They come from my own observations as well as from others in tech. These are the high-level buckets:

  • Overall direction
  • Growth
  • Entrypoint
  • Durability
  • Personal interest

Overall direction

Efficiency and change. Like the biological environment, the economic environment is generally efficient. Incumbents tend to be well-adapted and difficult to compete with head-to-head. When do new organisms and businesses emerge? Changes in the environment unlock opportunities for new entrants. When something changes, incumbents may no longer be well-adapted, allowing better-adapted (or at least nimbler, more adaptable) new players to gain a foothold. When exploring directions, ask: why now? What political, social, technical, or economic changes make this business newly possible in a way that wasn’t 10 years ago? 5 years ago? If there’s no good answer, then it’s likely either the opportunity is bad, or a competitive market already exists.

Step 1 and step 2. Successful companies usually start by dominating a small market (step 1) and then scaling (step 2). Step 1 often means solving an acute problem for a small, mostly homogeneous group. Then step 2 is when either that initial group becomes bigger, and/or you’re able to solve more common problems. For example, in 2008 if you’d solved problems for iPhone users, your natural step 2 would be to ride the wave of smartphone adoption as the number of iPhone users grew year-over-year. Oftentimes the shape of step 2 is not obvious starting out, but you can consciously pick a space that’s likely support a step 2. Ideally, step 1 results in a sustaining revenue stream that lets you ride out whatever secular trends lead to step 2. Step 1 can be thought of as the “beachhead” or “entrypoint”.

“Moving away from” and “moving towards”. Successful startups make something people want. There are two approaches here: (1) automate away the things people don’t want to do and (2) give people more of whatever it is they actually want. These are related and often conflated. Startups often offer both, and even those exclusively focused on #1 tend to explain their value in terms of #2. (“We free up your time to let you do what you love.”) Still, this can be a useful dichotomy. A whole class of services, including AWS, Gecko Robots, Task Rabbit, and Stripe are focused on automating the backend of society. These companies optimize for efficiency. Others like Facebook, Airbnb, Snapchat, Oculus, and Youtube give people more of what they want with all their freed-up brain cycles. They optimize for attention.

Aggregation theory. Ben Thompson’s theory explains how the internet shifts value in the supply chain. Previously, distributors competed on exclusive relationships with suppliers. Now, aggregators like Uber and Airbnb own the consumer relationships and commoditize suppliers. Initial advantages are created by building a beachhead with users through a superior user experience, and moats are built by the subsequent lock-in of suppliers eager to serve those users.

Growth

Standardization and universality. There are a handful of industries where a small number of people can have outsized impact: technology, finance, media, transportation, energy, etc. The key properties of these industries are standardization and universality. Billions of people need the things these industries produce, and there are standard interfaces that allow a single team to create something that can be used across the industry. Standards include things like computing platforms, the stock exchange, the automobile, the TV, the oil barrel, technology protocols, movie screen, shipping container, etc. In industries without these properties, scale is still possible, but it requires the company to create the standardization instead of riding existing ones. The absence of standardization is in itself an opportunity to create value (see Clever and Flexport), though that usually comes as step 2 rather than step 1.

“Technology” plus “way of thinking”. Impactful startups often bundle a novel way of thinking with a technology that complements that way of thinking. They succeed by spreading both through the world. The way of thinking usually seems weird at first, but eventually becomes universal. Facebook’s novel idea was putting your real name and private information on the web. Airbnb is sleeping on a stranger’s sofa. Slack’s is using emojis at work, among other things. Peter Thiel describes this as: a good startup is a conspiracy to change the world. Invariably, there will be existing players in your market. You should have a hypothesis about why they’re thinking about it wrong if you want to compete.

Bits vs atoms. Bits are easier to work with than atoms. Companies that can be reduced to shuffling bits are attractive, even if the bits are regulated. Turning atoms into bits unlocks value in this way. There are several approaches:

  • Replace atoms with bits. Netflix’s replacement of DVD shipping with video streaming replaced an entire set of atoms with equivalent bits.
  • Make atoms as easy to change as bits. Amazon’s factory automation makes moving bits as easy as moving atoms.
  • Measure atoms as bits. Samsara’s sensors make monitoring physical industrial processes as easy as monitoring bits. As bits, these processes are easier to manage and automate.

Distribution. People only spend so much time each day looking at pixels. It’s hard for them to pay attention to anything. Why would anyone use your software to solve their problem instead of whatever their current approach is? This requires some creativity. Lemonade uses a novel financing model that rewards non-profits for selling their insurance products.

Entrypoint

Superpower. For someone to use your product instead of whatever they’re doing now, it needs to given them a “superpower” (10x better). Inertia is massive. Looker gives marketing people the super power to analyze user data without having to code. Segment gives them the ability get data without having to go through engineering. Stripes gives developers the ability to accept credit cards. Clearbit gives salespeople the ability to get otherwise secret metadata on possible leads.

Revenue-growing vs cost-cutting. There are two ways to create value with enterprise software: increase revenue or decrease costs. It’s much easier to sell software to revenue drivers (marketing, sales) than to cost centers (finance, accounting, legal). Companies are more than happy to give you $1 if they get $3. The best situation is when they only pay on incremental revenue that you drive. AdWords, Stripe, and Salesforce are appropriately aligned with customer incentives in this regard. Keep in mind that some things that look like costs like compliance actually directly drive revenue (if they eg help sell to enterprise clients).

Vitamins vs painkillers. Most people don’t take their vitamins, but people in pain definitely take painkillers. Product pitches that elicit the response “oh, that would probably be nice” are vitamins; pitches that elicit the response “how much can I pay you to take care of that today?” are pain pills. Step 1 should be a pain pill. Certainly some people take vitamins, but in general people either don’t care or forget solutions to problems that aren’t actually painful.

Single-player and multiplayer. If your product requires multiple users be valuable, it will be harder to gain early adoption. This is especially true for enterprise software. If a purchasing decision-maker within the organization has to convince their coworkers to adopt your tool for it to work, they’ll shirk away. Their reputation will be on the line.

Lower an existing cost / improve an existing experience. A great pitch to a potential client is: you already pay for this. We’ll do it for cheaper and give it to you with a better user experience.

Durability

Recurring revenue. The best businesses create and capture value on an ongoing basis. Once acquired, a customer relationship should continue indefinitely by default, and at least lead to low-friction repeat sales. Without this attribute, it’s harder for the business to become durable over time. The team needs to be in constant sales mode finding and pitching new customers. Also, recurring revenue is nice because it aligns incentives with the customer — you don’t want to optimize for selling something and then skipping town. Subscription services are appealing for these reasons.

Consumer brand or unique IP. Most modern products have long supply chains. Some pieces of the supply chain create and capture more value than others. For example, you’d rather be Google than Comcast. Google needs Comcast’s pipes to deliver its services, but it’s a lot more valuable. Similarly, you’d rather be Uber than a road maintenance company. The highest value parts of the supply chain tend be the end-consumer relationship and proprietary hard-tech intellectual property. The best companies have both. Hat tip: Ben Thompson.

Low cost to serve. Businesses that require operations people to take action on every transaction/sale… these businesses are hard to scale. Ideally you have revenue that recurs without a human needing to be involved all the time. Samsara vs Flexport. Samsara takes effort to get a new customer up and running. But once they’re up and running, they keep paying every month even if they don’t require human attention from Samsara. Versus Flexport does.

Rocks vs sand. If you need to fill a jar with rocks and sand, put the rocks in first, then the sand. Better to start with the big stuff first, get those done and derisked, then focus on the small things. So it is with prioritization in any complex project.

Personal interest

Swing and a miss. When considering a project, ask: are you interested enough to be happy swinging and missing? If the answer is yes, your upside is unbounded and your downside is bounded, since you learned about something you’re interested in. If the answer is no, you should find something else, because you’re almost certain to miss, or even worse, find yourself in a situation where you’re not sure whether you’ve hit or not, unsure whether to proceed working on something that you don’t find intrinsically rewarding.

Who do you want your customers to be? You’ll be spending a lot of time with your customers. This is true even if your sales process is self-service, and it will be especially true if your sales process is hands-on. Who do you want to be spending a lot of time with? Empathizing with? Learning from?

Excitement. Rigorously applying all these frameworks is paralyzing. You can’t anticipate all the things that will happen or guarantee success. You just have to get excited about something and dive in!

--

--