PULL is everywhere

FAQ #1 about PULL

This week’s podcast:

PULL is, I believe, the most important concept in startups and business. It is the hidden load-bearing wall that explains why some products take off but others don’t, why customers really buy things, and what is the “shape” of a new thing is needed in the market. - The PULL Framework, Refined

When people are trying to understand the PULL framework, they often ask very similar questions. Some of the FAQs on my backlog are:

1. Does the PULL framework apply to enterprise? B2C? Deep tech? Hardware?

2. Do you need PULL in order to succeed?

3. Does the PULL framework assume buyers are rational / logical / premeditated?

4. Does the PULL framework explain the emergence of new categories of products, like Steve Jobs and the iPhone? Henry Ford and the automobile?

5. What does PULL feel like in a sales call?

If you have others, send them my way!

In today’s post, I’m going to focus on the question, “where does the PULL framework apply and where does it not apply?” I get this question for two reasons:

1. My “canvas,” as it were, is mostly B2B AI startups. It’s reasonable to wonder if the framework + approach applies elsewhere.

2. Everybody secretly hopes it does *not* to apply to them. (I wish it didn’t apply to me! I wish I could just build magical products without selling and the world would shower me with adulation + money)

Background

If you’re new, here’s the most comprehensive post I’ve written on the PULL framework. In short, the PULL framework says that a buyer will rip your product out of your hands if these four things are true (while everybody else would be weird to buy your product):

1. There is the project on a potential customer’s to-do list, which is being prioritized right now.

2. There is a reason the project is urgent or unavoidable now, versus all the other projects they could prioritize on their to-do list’s infinite backlog. They are in some sort of situation where they would be weird if they didn’t prioritize this project!

3. They consider a list of options for accomplishing this project they’ve prioritized. These options are not just our direct competitors - they might be other technologies, hiring, or a variety of other things.

4. But! The options they consider have severe limitations that prevent the potential customer from making the kind of progress they’re trying to make.

The whole point of a startup is to find repeatable PULL, which causes fast growth. PULL, in other words, defines a buyer’s unmet need - which pulls companies into existence and makes products relevant!

Let’s look at how PULL applies in enterprise, consumer, hardware, and deep tech.

PULL applies in enterprise

Big enterprise deals can’t be governed by something so simple as PULL, can they?

A friend spent two years hitting his head against the wall trying to sell an AI executive coach. He tried selling into HR departments, coaching firms, and to individual execs with basically no traction.

One day, he had a random interaction with an exec at a big healthcare company.¹ He had no healthcare experience or knowledge, but he asked the exec a very open-ended question, basically: “What’s at the top of your priority list that you need help with?” The exec responded that she wanted to understand what people were saying on social media about certain diseases.

The founder said, “Oh, we may be able to help you with that.” They scheduled a second call, where the founder pulled together a sample PDF report with some of the market intelligence the exec had asked for. The exec had some feedback but generally liked the direction, and they worked together to craft a contract. At the same time, the founder was scheduling conversations with other healthcare execs to see if they also had this on their to-do list and wanted similar reports. Turns out, they generally had similar demand. Each of his early contracts - all enterprise, all healthcare, all pre-product - closed in 4-6 weeks.

This might sound like a lucky break, but almost every startup story (enterprise or not) goes like this.

Need more examples? A university team got their first customer - a Fortune 100 entertainment company - from random outbound pre-product, turned around a vibe coded demo in a few days and it was a good enough “fit” for the buyer’s demand and got the contract done in a few weeks. A conference organizer sold AI governance solutions (a mix of product + service) into the biggest automotive companies to bootstrap her startup.

All this to say: PULL is extremely important in enterprise! Because it’s so difficult to buy things in enterprise, your champion (the person with the project on their to-do list) MUST have something they REALLY NEED to accomplish. The project on their to-do list is what’s going to cause them to suffer through procurement and 47 group demos; they are trying to make something happen, and believe you can help them!

I will say, there is a nuance in enterprise: The project on the champion’s to-do list might be somewhat idiosyncratic per customer. John wants to demonstrate he’s using AI to get promoted, and thinks that your AI tool will help him do that. Jenny has been given the unfortunate mandate to shrink her team by 30% while doing the same amount of work, and can only accomplish that with AI automation. John and Jenny both buy - from slightly different roles at similar companies with somewhat different projects. This is still “repeatable PULL”, it’s just not perfectly identical like you might get in SMB.

(This nuance can lead to a clunky product that’s defined by the sales team, but it doesn’t have to.)

One last example: Enterprise services. Consulting firms like McKinsey offer extremely variable services - from strategy to transformations to …other stuff… - for different buyers in different industries. Each project fits each buyer’s idiosyncratic PULL. But the reason McKinsey has been able to consistently deliver bespoke/idiosyncratic services at scale is because they have built a “factory” that standardizes everything other than the idiosyncratic project content. Seriously, McKinsey has standardized processes for everything: recruiting, training, scoping projects, kicking off projects, slide formatting… all so that they can consistently deliver on the specific thing the buyer needs at a very high quality.

PULL applies in consumer

Using the PULL framework to identify an ICP and a need in consumer is pretty uncontroversial. An example: Rob Carpenter, my podcast conversation partner, runs Cumbre Media - a media company that builds STEM-focused content for families. They have two sides of their business - a B2B side (sponsorships / co-production) and the media side, which is obviously B2C. He talks about how he uses the PULL framework on the B2C side to (1) articulate needs to identify new media properties to build, and (2) clarify exactly what their existing products are, who they serve, and where they “fit” in the market.

The controversial part is 1:1 sales in consumer. I think it still applies.

Take, for example, Sarah Blakeley from Spanx: She went and sold Spanx in malls herself in the early days.

Take, for example, Howard Schulz from Starbucks: He spent all his time in the first Starbucks store unfolding it from a fancy Italian espresso bar to something that fit the American yuppie’s PULL. He did this by watching where people got stuck

A smaller example: I wrote about the small American Oak distillery back in 2021. The owner had just opened a tasting room, and was sitting out on the front porch listening to people walking by who were considering coming into the tasting room. He heard people turn away after realizing it was just whiskey - so he changed the name of the tasting room from “American Oak Whiskey Tasting Room” to “The Oak Room” and started offering other drinks. Others wanted food - so he started offering small plates. The point is he did this by watching actual and potential customers when they were considering a purchase decision. Four years later, they’re still rocking 4.9 stars on Google as the top-rated spot in town.

In all of these cases, 1:1 sales and customer interactions do not scale, obviously. But they help you learn what actually works - what fits PULL. You can only scale something after it actually works.

PULL applies in hardware

Surely you can’t sell-then-build in hardware? Wait, Kickstarter suggests otherwise. That’s B2C hardware, what about B2B? I’ve seen nonexistent hardware pre-purchased for delivery 6+ months out by serious businesses off a 1-page overview from a hardware startup team. Moving on.

PULL applies in deep tech (sometimes)

I have worked with deep tech focused teams out of Harvard and MIT. There are two kinds of deep tech projects:

1. Fundamental research with a specific research goal

2. Research geared towards commercialization

The former is not a company, though may become a company through some happy accident - they invent something in a lab that happens to fit PULL. For fundamental research, PULL does not directly apply, though some fundamental research organizations have a research goal that’s loosely tied to some sense of a need. OpenAI’s Goal 3 from 2016 directed their research towards a chat agent. Pharma research is a great example, too: They seem to do research with some market need in mind, though they almost always find that their invention has nothing to do with the market need they set out to solve.

The latter case - research directed towards commercialization - is common, and PULL applies here - we need to find a need, and we should try to pre-sell because of the long lead time of research. Unfortunately, these researchers tend to go through university commercialization accelerators… where they are often told to interview customers, not to sell, because why not waste your life as a postdoc in a research lab making $0.75 per hour while building something that you think is validated but actually isn’t. Sigh.

PULL applies in every moment

I can write a longer post about this if you’re interested - but the concept of the “to-do list” is a foundational model for understanding human action. In each moment, you take action (or no action), and your action reflects the to-do list in your soul. The generality of the framework is by design: We need to understand the source code of human action in order to truly know how to make our cool little B2B AI companies take off.

—

PS: A few VCs have asked me to offer a “sales pitch sprint” to their portfolios: I review your recorded sales calls and craft a new script + pitch + demo for your real ICP, then give you feedback on how you execute the new script & teach you how to iterate it. If you’re interested in something like that, email me at rob@reframeb2b.com. I’m testing it out with a few startups (paid, but less than what it will cost in the future). Or if you want my help as an advisor for a longer period, learn more HERE.

1. Slightly anonymized