Beware of step 2 to step 3. I call it "chasing the whale". The problem is the "whale company" knows they are a whale. You'll slowly watch your entire development timeline and features shift to benefit only them. Your company basically becomes a contractor for them to get X feature working (and nobody sees this feature being helpful to any other customer). You can't get rid of them because it looks terrible for funding that you dumped such big potential. Technical debt piles up from priority shift and then the whale is gone. You were only a POC for them, and in the wake of being side lined as a priority other customers leave.
I've also worked at consultancies who let their software product strategy be effectively doomed by this.
The sales team were always underselling the software (often just giving it away for free to land consultancy contracts); promising the moon to "large" companies because it looked amazing to get such a name on our homepage and looked good at the monthly company meeting.
But the reality of large companies is that they are a many-headed beast, and just because you've sold some software to "BigCompany", doesn't actually mean much if you've actually just sold to the 4 person HR department of a local branch of a sub-division spin-off.
That department leverages their "BigCo" name to get the world at a knock-down price, while the small company is throwing everything at chasing the promise of huge future revenues "when BigCo roll out fully" which never materialises, or happens in name only.
Very often it turns out it's only ever "Dave" that had even heard of your small company / product and was just using it for some other BigCo political games. As soon as he leaves BigCo, because he leveraged "his HR transformation project" into a better job offer elsewhere, you struggle to even get a single person at BigCo to pick up the phone. To be fair, they'll happily keep paying your invoices for the contract, but remember you sold at a loss for the promise of future revenue which is now never coming.
But hey, it's okay, because your own sales team have also moved on and are now promising that if you just rush in feature Y then we're sure to get a sale with MegaCorp. ( Repeat endlessly, to the stress of the rushed development team whose time has just been undersold again. )
This is accurate. We even underpriced the consulting and custom work. Didn't charge for project management, Product management and rarely design. Only for Dev and QA.
Every weekly the sales team had the same bigcos "in the pipeline" as six months ago and they weren't replying to emails. But you have to help us get that SOW ready by tomorrow so we can send it to them.
I work for a multi-billion dollar publicly traded company and I've watched this happen to a new product that our org launched last year. One big customer set the direction early on and got used to us building exactly what they asked for. As we've gained more customers we're having to tell them no more and more often and they're getting less and less happy. We're locked in to a long contract so they're not leaving anytime soon, but it's turning into a constant battle with them.
Or, you get past the POC and “land” the whale. You take a deep discount on price to lock in the revenue. Now you have revenue, but insufficient to hire and grow. Rinse and repeat. All the debt (technical and financial) and a dim future.
This happened to me and my company with a F10 retailer in the US long long ago
The reality is, most times you have no options and that whale is your highest likelihood exit
The reality is, this is what being in a market with what are effectively monopolies looks like. You either bully your way to the top or have to deal with being bullied till you can get out.
This is why you need to actually say no to giant enterprise clients until your company is ready enough to weather the multi-year sales cycle with a strong revenue base outside of it, even if those enterprise parties are interested. I think a lot of B2B executives forget this lesson. A sales cycle your company is not ready for could kill it if attempted prematurely, even if it could theoretically close.
It just means spending 12-24 months of time, from everyone including ceo, vps, Product, cto, devs, designers, marketers, to get th contract signed. 24 months of these people's wages is a lot.
> There are all these horror stories in Silicon Valley of these companies who got to 150 or 200 million revenue and then didn’t grow
Yeah, what a nightmare that must be.
I’m being a little glib here — I understand that if you’re making $200M at a cost of $400M and your investors are expecting $2B and you stop growing… you’re in deep shit.
But christ, people, can’t we get a little perspective?
"Infinite Growth or Death" is an impossible goal but still people seem to strive for it.
Though most VC funding seems to have the explicit goal of "Burn cash to drive out competition so you have a monopoly, then abuse your market position to wring your locked-in customers dry", which feels awful from a consumer POV.
It sounds like you already have exactly the right perspective.
It _is_ a nightmare. That $200M isn't their money, it's not in their pockets, and it's far short of what it was hoped to be. You framed it perfectly, and so did they: they're in deep shit, and it's a horror story.
What do you mean can't we get a little perspective? In terms of running a company vs the real world? Sure but this thread here is a more targeted conversation.
If you make 150-200 M in revenue and can't grow bridge the gap from revenue to profit to sustain operations you have a serious problem on your hands. I think you need a bit of perspective on what that means as you now have a lot of dependent workforce that you likely have to layoff to bridge it.
Yes it is impressive to make it to that point of revenue but your future is looking difficult especially as you probably have layoffs in the future - which is awful.
VCs want you to go big or go home so they want you to hire. Hiring decreases your runway dramatically.
$100 million lasts practically forever for 2 people. $100 million lasts through most problems for 10-15 people. $100 million doesn't last very long at all for 100 people.
If you have $100m, the implication is that the VCs want to see a unicorn (>$1 billion). You will be engaging in a lot of sales, marketing and engineering activity to get there. Your activity is expensive.
And nobody is getting $100 million anymore. Good raises are now in the $10 million range and that doesn't last long at all even if you only have a small number of employees.
It’s a nice meme about Twitter being an atrocious business etc (and it wasn’t great/it was struggling for much of its existence), but Twitter made $1.2 and $1.4 billion in profit in 2018 and 2019 respectively.
Utterly pointless because he didn't interview failed startups about how they felt.
It's like doing an experiment and only counting the "successful" results - this is worse than "bad science", it's intentionally misleading science that actually sets us back.
We really don't need any more survivor bias. Please go interview some failed startup founders. There might have been a crucial difference in step 2 of these 5 steps that successful founders did and failed founders didn't, but we won't know about it because there's no data from failed founders.
I guess if the point is to find out How long it took different companies to find product-market fit (title of the submission) you could argue that failed startups never found it?
I agree with OP - lots of survivorship bias here. To be fair to the author - going through the heaps of companies that didn't make it would be a sig more work.
Most likely the failed companies didn't find it but possibly were close or other issues sank the company. Curious actually if there are companies that found product market fit and didn't survive. That's an interesting group.
> Achieving PMF is just one stage of a startup's journey. Once PMF is achieved, the company still needs to scale, navigate competition, manage its finances well, adapt to changing market conditions, etc. Mistakes in any of these areas could lead to failure even with a strong PMF.
> For example, one of the first successful smartwatches, Pebble had a great product-market fit with a passionate user base. Yet, they struggled against larger competitors like Apple and Fitbit and eventually had to sell their IP and shut down.
I think the point is that PMF is a significant milestone on the journey. But it's not the destination. You can still mess it up despite achieving PMF.
And I don't think "keeping investors happy" is the key, either. Plenty of successful founders with unhappy investors.
I'm struggling to see how you could force adoption of a product in a way that matters. The example given was Pebble, who were struggling against Apple. Apple still needed PMF, despite their market dominance. If their product was so bad that it didn't solve the market need, then it would have failed despite their money and effective monopoly. I don't think PMF is irrelevant, it's just not the only necessary thing. You can still fail despite having PMF, but you have to have PMF to succeed.
But I do use them, so I must gain some benefit from them.
There's even more trash that doesn't get used and we never hear about (or hear about and decide not to use). A lot of that is produced by large companies with big budgets.
Anyone else a bit skeptical of the “time to having a live product” timeline of many of these companies? Maybe I’m wrong, but e.g. there’s no way Gusto produced a working managed payroll product in a month. There must have been either been significant pieces of their product that were pre-existing or else they built an interface to someone else’s product.
I love me some Kent Beck, and I know your comment was tongue-in-cheek, but I doubt he was interviewed for this article. 1 month for that kind of a product is short enough that it’s either a clerical error or someone is BSing big time.
Great inspiring article, but I think there’s much more nuance in reality that it doesn’t address.
My company (we’re multi-product) has been working on a new product that ticks a lot of the boxes that are proposed here, but that is still nowhere near PMF.
There are > 100 paying customers who pay ~$30k for the product.
There is a small set of these customers who absolutely love the product.
That’s steps 1-3 done.
But…
Overall retention is low (varies between 55-60%).
Sales are pretty flat and seem like a real struggle. Customer acquisition cost is 130-150%.
Just because you got these first 3 milestones doesn’t mean that the rest naturally follow.
My view on this product stalling is that it doesn’t really get the results that the customers think it’s going to and they cool on the idea and stop using it, but… you have companies like Snowflake that went to market claiming to be the one data platform to rule them all, which was also a large over-promise (it’s more expensive, slower, complicated than what they promised) but it grew explosively and IPOed.
Even towards the end of the journey that this article describes there’s still major rolls of the dice that affect the outcome.
> But… Overall retention is low (varies between 55-60%).
Assuming you have a yearly terms (which you should, if at all possible), try to identify customers who might churn (using the product less, etc.) and spend some time with them -- literally, just send them an email to hop on a call. Try to understand why they are using the product less. Do the same with the ones who love you. Figure out what is different.
Usually, this will grow into a "customer success" role, but it has to start somewhere.
Side Note for Lenny, in case he is reading: whenever I try to add my email, I get "We were unable to validate your email domain". Of course my email is valid. I just use a forwarder email address, mixed with a catch all, so that I can do lenny.newslette@myexampledomain.com. But your website would not accept it. It is the only ever website that rejects it. Of course I would never give my real email.
If you're using escribime dot com (as per your profile page), you're missing an MX record. Software will usually default to the A record in this case but that's not exactly standard.
This is such a great post for so many reasons, but I'm glad Lenny made that chart in particular. The overnight success fantasy is toxic and I've seen it infect (and sometimes ruin) many startups.
All startup advice is either tautological or so ambiguous as to be as true as its opposite. The reason is that no one actually knows how to do this stuff. No one has identified what ingredient or set of ingredients distinguishes successes from failures, except for rapidly learning and pivoting (but but of course you can’t pivot too quickly, as this graph shows!)
Something feels off about that and the other time lines. I’m sure I remember paying for Airtable pretty soon after it launched for a couple of months. Perhaps the timeline chart is literally just for the B2B channel?
I don't think the data is very "scientific", e.g: slack was officially launched in 2014 and got immediate traction. Yes there were 4 years of development but it wasn't even the product they wanted to sell.
If it takes a year or more to go from launch to first paying customer, what are you doing during that time?
Of course, you could be purposefully giving your product away while building the premium, paid-for features. I think that applies to some of these companies, and I'd be more curious about the time from launching that version to paid customers.
> Stop thinking of product-market fit as a yes or no question—but instead as a process of finding fit with more segments of the market.
This seems really true to me. Though the nature of the search for PMF with more segments changes after passing some milestones. It’s more difficult to make large changes in the product once you have a basic level of traction and growth.
Cool viz, yet I find the timeline misleading for some companies. Taking segment for example, sure, the moment they pivoted to CDP kind of business they reached PmF quite fast. But that happened only a year after their initial idea. They kept building and failing until they finally got the segment idea and famous Show HN.
What an excellent post. I love the "first felt PMF" criterion which is the right way to think about it because, like validated by interviews in this blog, most of the times you remain unsure that you have actually found PMF... and that's perfectly ok!
Notice that products get more amazing towards the bottom of the chart.
Building a great product is the challenge. Once you choose target customer and problem to solve, PMF is trivial but the hard work is still ahead. iPhone, Falcon 9, Starlink - none of these ever pondered PMF. They set out to build the most batshit amazing product for a specific purpose, and you can add Figma to that list
The whole concept of PMF seems to be based on the lottery-ticket view of startups, "I don't know who is the customer or what they need, but someday I will randomly stumble across it".
> They initially thought they would be providing search to companies for their intranet.
Really? Page and Brin were focused on public web pages and links from before Google existed, through their PhD research [1], so switching to intranet seems strange.
Great article. I would like to also know how close, or the closest time each of these companies felt to dying, or if they ever thought they wouldn't find PMF. Or when they knew they would get to PMF. Or if they always believed they would get to PMF.
Is a dead company just the same as one that never found PMF? Is that how survivorship bias is accounted for?
Wow, the s curve for "first felt PMF" around 2 years really drives home Dan Luu's observation that discontinuities are usually there because of some externality
> One of the most interesting takeaways is how often founders never fully felt PMF
Perhaps it's a shared trait among great makers. Like an industrious person feels their work is never done, a good maker feels they've never fully achieved PMF.
It is a cool visualization. I really wonder if the classic flaw of "what even is PMF" goes to show here. You can somewhat see that once you get past the first few companies.
I've seen this happen multiple times.