Are these really "mistakes" or are they just the random process of value discovery, which naturally involves a lot of failures. Finding product-market fit is not something you can engineer with perfect foresight. You raise money, you build something, and then you have a limited time to see whether your hypothesis about the market was correct. Often, if not most of the time, you fail to get the fit you need in order to succeed.
There are definitely execution failures, but in my experience spanning 20 years, I have seen far more failures that are really just failed experiments rather than execution failures. A half-decent team can get product-market fit; once you're there, attracting the people who won't fuck up the next phase is relatively easy. If you don't have product-market fit, no amount of genius will save your company.
> You raise money, you build something, and then you have a limited time to see whether your hypothesis about the market was correct.
Or, you get to know an industry, see a need, design a solution, bounce your ideas off people in the industry, do some mock-ups to get some initial acceptance of the idea, ask if they'd pay for it, and if/when all that comes together, then start to build. If you cannot build it with your available resources, maybe look for some funding.
And then once you know an industry, see a need, etc... "You raise money, you build something, and then you have a limited time to see whether your hypothesis about the market was correct."
Works well if your startup is filling in a niche within an existing industry. Does not work well if your startup is creating a new industry. There is a reason why the failure rate for the latter situation is abysmal. But that is also the reason why the returns are so large and sought-after for that situation.
You also dont see this mentioned that often, but failed attempts can be key factors of success to the successfull ones that comes later.
Therefore, we should be less intolerant with failure and understand that it forms some sort of a ladder where the failed attempts are steps that lead to the successful ones.
I miss this sort of "tech archeology", and i think that the status-quo of tech would do a greate service to us all if they aknowledged the failed attempts more, as this would give more incentives to people to take risks on innovation, knowing that even if they fail, their efforts would be remembered as it would lead to something that worked fine, with a couple of tunings.
There are a lot of bold and bright people that did amazing things that may (or may not) lead to big things, but that face the cruel reality of not being remembered because they are not the success case.
Because in the end, we are only able to succed when we embrace failure and stop being afraid of it, knowing that even if we fail, it might be key to success if not ours, from others that can pick it up from where we stopped.
> Finding product-market fit is not something you can engineer with perfect foresight.
I think the larger problem is that is marketing is near ignored and not just by engineers either. Not many people want to talk to potential customers or users to see if there's product market fit, so they don't talk and listen to anyone outside of the friends and their team which is a huge mistake when they just go ahead and build. This is why I feel that Product Hunt's various hack fest events are more valuable than YC's virtual startup school. Product Hunt's hackfests are ingenious marketing classes in disguise. Product Hunt really forces you in a methodical way to market your product before it's built. What most people don't realize is that it's so important to build an audience before you finish building your product. Ryan Hoover and Co are right: "Build it and they will come" in most cases is a myth.
talking to potential customers and users about product market fit
is not a given it will work
Making a product based on your experience of a market - 25% chance it works out (of which 5% chance it will become a big company)
To that, even after talking to customers
you are only taking it to 30% chance it works out, and of which 7% chance it will become a big company
*
your hypothesis is based on data points
hopefully things like actual experience in the market
market research
seeing what sells
seeing what doesn't sell
*
Customers and users are just as likely to give you bad data (of what they THINK they will buy) or impractical data (something that benefits them and doesn't benefit you)
It only adds a few percentage points to your likelihood of success
on the other hand - seeing what customers actually BUY and SPEND ON, that has a lot of value
However, even that is a piece of the puzzle, and not the solution
With a startup nothing is a given. However talking to your potential customers increases the chances of success because it increases the chances of product market fit. It puts your hypothesis to the test.
Yes, those few percentage points count. Every little bit counts when the chance for failure is high.
I’m not really sure why you’re criticizing my argument.
This is entitled Software developers assuming they can jump into a completely different domain and instantly succeed. Sales? Fundraising? Business development? Operations? Logistics? Recruiting? Psh those losers make half of what I make I can do all their jobs at once no problem!
Oh wait this is hard and I’m not good at it I’m going to wonderful journey my butt back to a cushy ad company.
The process of value discovery involves all the business, ops, marketing, etc people you're talking about (they would be hired with capital or equity). No one is saying they can do it on their own.
-> You raise money, you build something, and then you have a limited time to see whether your hypothesis about the market was correct. Often, if not most of the time, you fail to get the fit you need in order to succeed.
A lot of people will try to counter it
However, that is REALLY what it comes down to
we have a HYPOTHESIS about the market. A product that we think will win/succeed/do well
The Market Always Wins
It either accepts your product (your hypothesis was correct, or became correct by the time you launch your product)
"By far the most common reason for shutdown was lack of product-market-fit which constituted more than half of the marketing mistakes"
During the dotcom boom of the late 90's the issues in the Valley were often attributed to a lack of engineering talent. However, after the bust many founders and execs realized that their failures were not due to a lack of product and engineering talent, but that there were too many engineers putting out products that the market did not want or need.
This study seems to validate that idea. Without a clear understanding of the market and how to reach your potential customers, your idea is likely not worth the time. Spend more time on understanding the market, and how to reach it, and you will have a much better chance at success.
I really hate those kind of content. Would you listen to someone who built a wall which collapsed ? nope
But you will tell me, if I can prevent the mistakes other made I will go further than them. Well I'm not so sure, for any mistake that someone thought they made there will be another person who identify this mistake as something which made them succeed.
These kind of articles are just porn for people who are not putting enough hours in working toward their ideas. (And top x things that made us successful is as stupid)
Sure the first 65% is work hard. But guess what? The last 35% is a much different ask that just does not succumb to that same hammer.
As the immediate parent implies, wall building is not all or nothing. Even in statistically controlled manufacturing processes there is std deviation. Not learning from mistakes is itself a kind of arrogance. In fact the preferred model is neither arrogant rejection or passive adoption. To use an a sort of Bhuddist analogy, we listen carefully and extract the gold from the gold+useless ore in the wisdom we hear and refashion it into our own art. This requires decent teachers (suppliers of know how) but also decent students who are interested in real learning not passive regurgitation. Making our own art is demonstration of mastery and of appropriate internalization. It avoids the silly extremes of man as an island and man as an empty if pious vessel to be filled with rhetoric that we can grandstand on.
Those of us in profitable and large organizations will see the same truths about market acceptance and customer approval in a modified way: often our organizations are too sterile and inward looking. Losing the customer starts slow but in that last quarter or two accelerates fast often with a profound finality to it.
Yeah I feel like these articles always end up coming up with apple pie positions. Like talking to customers more is good advice probably but it’s also the most common most stated advice for early startups.
These retrospective analyses, while intriguing to read, should be taken as more descriptive than prescriptive.
When you talk to someone who actually succeeded at something, they'll talk about platitudes like working hard, product market fit, iterating, listening to customers, raising a lot of money or not raising too much money, and so on. Many people go so far as to try to emulate the behaviors of their "hero" (Jobs, Gates, Page/Brin, Zuck, ...), but that's a fools errand.
In a similar way, these types of postmortems should not be taken as prescriptive. They too are full of platitudes.
I have worked for some companies that were total failures and some that were global successes. Having been through the failures first, I was excited to see how the magic worked at a massively successful company, but to be honest the differences are more subtle than anyone would believe. The big successes were happy to do things that had been disastrous at other companies in my experience, and did not pay much attention to the things that the popular press considered the secrets of their success.
As disappointing as it is to think this way, it’s much like explaining the weather or the stock market. I think people need a story, but both inside and outside the company we are just guessing why a company fails or succeeds.
Andy Rachleff, known for his articulation of product/market fit, says a company with a lock on product/market fit can do almost everything else wrong and still succeed.
Google firing all of its managers early on in its existence is a great example of this.
There is both a lot of survivorship bias and probably misattribution of what leads to failures and successes. I imagine a lot of it too is really indescribable, in the sense that a professional basketball player may write a blog post about how to play top-tier basketball, but reading won’t actually make me any better at basketball, nor will it give me the character/physical traits I need to be talented.
Good read with some really interesting information, but I don't know that I agree with the way it implies that there's a single, defining reason for failure. All of these things are intertwined - you can't get raise money if you don't have product market fit, and you can't bootstrap things if you aren't bringing in money. Team problems are tied into everything happening in the company. Tech problems are very much related to product market fit.
Like I said - good information, but I guess it's really the pie chart that made me feel like they were oversimplifying things.
Yeah, it’s sampling from a distribution where every company has problems in every one of those spheres they are mentioning, except only one of these issues will be the last one to finally finish the startup.
So it’s more about sampling from what kind of issues are more prevalent in a startup rather than what kind of mistakes they will be making.
My experience consulting for many startups is that what is called "Value add" is a real problem at more than 2 or 3% of startups. The momentum of the status quo is really strong and only gets stronger the larger your target company. For example, getting a 5 person company with 4 engineers to try your new Dev tool is an order of magnitude easier than a company 10x that size which is an order of magnitude easier than a company 2x THAT size.
I remember digging into the source material used in this article, and contacting the founder, not Nico, but the other gentleman, can't remember his name, an amicable friendly chap. Anyway, I said to him, that a lot of the founders, although I respect their transparency and willingness to talk about unfortunately failing, they appear blinkered to the reasons for their failure. I won't mention the founders reply.
Looking at PMF in terms of:
1. The definition of PMF
2. How PMF is drummed into founders
3. The proposed criteria to evaluate PMF
4. The actual things to measure within the criteria
I won't pull a number out of my backside to state to what degree PMF wasn't the issue, but I will say PMF was not the problem in some of the startups mentioned in the source material. In some cases PMF was not something they should have even been looking for at the stage they were at.
Another reason that cropped up quite a few times. Bad Marketing. Putting growth over retention, automation and efficiency over learning and personally figuring out how to sell in person isn't bad marketing to me. That's a bad decision.
1. Focusing on growth instead of retention in the early stages and certainly before PMF or at least strong indications of PMF, is a bad decision.
2. Focusing on automation and attempts at operational efficiencies when it comes to marketing and sales, again before PMF or at least strong indications of PMF is a bad decision.
When it comes to growth, automation and attempts at operational efficiencies, you do that when there are strong signals. You don't cut out selling in person, till you have an acceptable closing rate. Even then, I would still recommend frequent selling in person to learn more until there are diminishing returns.
Instead, you'll find an inordinate number of founders attempting to attain double-digit growth while there is triple digit churn. Automation of certain things like scraping lead lists, when the lists are nothing more than dump lists.
Hacking things out, when what they should learn is discarded because the process towards the end-goal cuts out what they need to learn, take on-board and become intrinsic.
Throwing up websites that are clones of well established competitors. Parroting without any deep understanding of why those things are on the site, of why those things need to be said. Oblivious to why they themselves are not in a position to say those things.
I would really like to say more on this. I would call it a pet-peeve, but I got a zoo full of them. If you want to carry on the discussion in private, happy to do so, but for now, I got work to do.
This. So many founders are enamored with their titles and their ideas more than actually building a real company. IMHO its the most obvious way to spot a future failure along with the amount of fluff PR and "thought leading" blogging...
CEO. The product owner should be finding what the customer needs, but as soon as you have more than one customer you'll have antagonistic pressures on what to implement next. It's up to the CEO to make the decision which customer needs to actually fulfil (with information and input from everyone else). It's up to the CTO to make those decisions happen in a way that doesn't stop the company doing more things in the future.
CMO. It's the result of the first decision the CMO makes (market strategy).
In short - CMO, CEO, COO AND CFO set the overarching strategic objectives (2 or 3 generally), then that gets turned into a framework for strategy and decision making to enable actions to be evaluated, enacted and measured.
At the highest level, the framework consists of:
1. Strategic aspirations - what are you aiming to do, and how will you judge the success or failure of those things? Goals, KPIs and success/failure criteria.
2. Where to compete (product category areas, pricing, distribution methods, experiential factors, communications and branding)
3. How to win - what will preferentially differentiate the offering, and how will it be distinctive against the market competitors.
4. Capabilities audit - what needs to be changed about what the organisation has/does in order to enable that, and how will the effects of that play out throughout the rest of the organisation?
5. Change management - how are those things going to be introduced/amended/reduced/removed and where does the responsibility lay for those changes?
Everything the organisation does should be in support of the strategic objectives, and thought about in relation to that framework.
As a result, the product-market fit comes as a result of product decisions, and how's it's positioned to the market is a result of the decisions for 2 and 3.
Note that product-market fit isn't just about what the product does, but how what it does is communicated to the market. Both of those are vitally important, bit people tend to forget the comms part.
If the org isn't big enough to have a CMO yet, then generally it devolves to the CEO.
It's why start-ups tend to have crappy product-market fit - CEOs tend not to be great at understanding marketing.
CEO is the primary but everyone in a startup needs to take responsibility for this. The core goal of a startup is to discover a "scalable and repeatable" business model.
We didn't find any significant correlation between funding and cause of failure. I guess, however, that this is because our data set was quite limited.
This was a surprisingly good read and data-driven.
I found it surprising that the top reason for failure was product-market fit followed by team. I would have thought finance (#3) and tech (#4) would rank higher.
Tech barely matters. If nobody wants your product, the right tech is a mock-up.
If people want your product, but you can't deliver because of tech, that might be promising enough to get funding for a better team. The dependency on a 3rd party product that's pulled seems real and dangerous though; you can't have security if you're built on something you don't control and can't replace under pressure.
Amazon is littered with awesome tech that took 2-3 years to build and launch but failed because there was no market for it. I worked on one such product, a micro payments web service.
It took me a while to realize that that’s just how Amazon operates. Carry out experiments all the time, most fail, but the ones who click take off spectacularly. As Bezos says, don’t take Hail Mary bet. Take smaller bets all the time.
The thing that just continues to strike me since I left Amazon is how many start-up ideas are buried within that company. Coming from logistics, there is basically no internal tool or process that couldn't be turned into a start-up. Quite funny actually.
I think finance and product-market fit go hand in hand, and it's a question of the founder's perception of which is more important. If you don't have a good product, you won't make any money, so for those startups it really depends whether the founder saw running out of money as the final blow or the fact that the product didn't find its market to be the overarching cause of things.
I also enjoyed it but I'm not sure I'd call it "data-driven" when the author puts "data" in scare-quotes:
> Disclaimer: the interviews are qualitative, even though the founders were asked similar questions, they weren’t asked to fill out a survey, so to gather the “data” I had to interpret their words and put them into boxes.
The thing about tech problems is that they don't show up as tech problems unless they are really bad. They usually show up in lack of product-market fit, lack of adoption, high turnover, and difficulty with sales and marketing.
For example if you chose Electron for your app and put no effort whatsoever into slimming down what you are running inside, users might decide they don't want to use your app because it uses too much CPU and RAM. If you use a slow database as your SaaS app/site's backend, your site will be slow and users might bounce during trial. If you pick an unwieldy, bug-prone, or weird (a.k.a. hard to hire) technology you might experience a slow development cycle and find it hard to build enough features or fix enough problems.
Most of the time people won't tell you these things. They'll just move on. Some of them are subconscious, like performance issues. The app just doesn't "feel" good, it's not a pleasure to use, etc.
If you are trying to sell shit on a stick, your problem may not be marketing... but you can always blame it on marketing! If your stack makes your development cycle slow, maybe you're not finding product-market fit because you're hiking through the woods with an elephant strapped to your back.
Good tech choices are fundament. They enable everything else to work well. You need everything else of course... good tech alone won't cut it. But the rest of your company's operations are going to flounder if they are not properly equipped or if the product sucks.
Are there any products that people really did want, for people who had the willingness and ability to pay for them, which are not in the market because of startup blunders? Put another way, other than being illegal, is there a problem that growth has not solved?
Not having concrete examples, but the book 'crossing the chasm' claims, that the majority wont buy from you whatever you have and they need unless they know 2 other companies (succesfully) using your product. Supposedly the step from a few sales to fully tapping the market is hard even if you have great products.
One of the big issues I always find with "Closed Lost Reasons"[1] like this is that it rarely captures the truth accurately. It misses the truth in a couple of ways:
1. Deliberate obfuscation by main actors who want to protect their jobs ("It is difficult to get a person to understand something, when their salary depends on their not understanding it.")
2. Continental drift- the farther you get from the action (or, more likely, the set of actions) that precipitated the event, the more you rely on memory and hearsay
3. "The Rage to Conclude"- humans LOVE denouement, to “know”, to have things wrapped up and make sense. But rarely is the world so simple. We try to boil a set of reasons down into one über-reason
So on the factory floor, the people doing the labeling are incentivized to shift the blame and obfuscate their role in the loss. And then the manager is trying to sort through all the noise to identify which set of reasons is the most likely for this loss. There's a lot lost in translation from the "truth" of how the deal was actually lost.
------------------
With this particular article, I categorize it in "content marketing babble": advice that is generalized to the point that it is not useful to me. The Common Lessons section leaves me with general advice but no tractable ways to understand what that looks like in the real world with real constraints and stresses.
IMO, the Lean: validation section could be extremely valuable IFF they supplied hard examples from the failed startup dataset that fueled this article. Like "Here are links to the case studies where the team behind [Closed Company A] felt that they could have spent more time validating before building. This is the featureset that they built prematurely, and here they ponder on ways they might have tested it cheaply"
------------------
All that aside:
I've actually been quite an avid reader of Failory for a couple of months now. I check in every so often to go over new interviews. I think the interviews (https://www.failory.com/interview-failure) are a lot more useful: the format allows for more nuance, more context over the conditions of the business, and more reflection.
I often find myself wanting to ask more questions in-line to the interviewee, kind of like a hyperspecific, morbid Quora question.
[1] I'm co-opting this term from Sales/Marketing lingo. I originally wrote Root Cause Analysis, but I know that RCA has a very specific software definition and set of rituals, so I wanted to
>Deliberate obfuscation by main actors who want to protect their jobs
Sure. There is mal-adaptive behavior, and it's not consequence free.
>rarely captures the truth accurately
Accurately? There's nothing going on in your reply to suggest you have distinguished insight to anything. Throwing out three more modalities of problems amounts to what-aboutism. You have attempted to re-contextualize the original write-up as disingenuous something we see a lot of pointless political wrangling.
The salient question is what we can agree on as to common failure modes, rather than that argue without intersection. No accounting can be complete, and beyond some level of specificity more detail is useless. We're looking for 20% of the causes that explain 80% of the bad outcomes. Think Chaos Report that had a good run in the 1990s.
Note that following the lean startup method does not itself guarantee success. I know at least a half dozen founders who have followed the Lean Startup method (several of whom were actually in Steve Blank's class at Berkeley), and in all cases the result of their 100 interviews was "Welp, nobody wants that product." Which I guess is preferable to investing 3 years and $15M in venture capital into it, but it also doesn't get you any closer to a successful startup. (When I followed the Lean Startup method myself, it was actually worse - our 100 interviews said "Yes, we definitely want it", we built it, and then nobody used it. At least I took every engineering shortcut possible to build & validate it quickly and didn't waste time making it bulletproof before people tried it out.)
Also interestingly, a large number of gigantic corporations today did not follow the Lean Startup method, and arguably weren't trying to found a startup at all. Apple, Google, Facebook, E-bay, Dropbox, AirBnB - they were founded because their founders were curious and trying to scratch their own itch, and it turned out lots of other people had that itch too.
If I had to distill lessons learned from about a decade in startups (7.5 years founding ones of my own and 2.5 years working for other people), it'd be "Trust my emotions more, be curious, and quit trying to found a damn startup." The problem is that you need to judge new ideas emotionally rather than rationally, but when you have the "I'm going to build something awesome and get totally rich" mentality, it distorts all your emotions. Better to get rich by working in finance or for a FAANG, reserve some time on the side to explore, and only set out on your own when there's something incredibly exciting on the table.
I dunno... if you've got first hand experience in 80 startups then maybe their failures are because you didn't spend enough time in any single one of them.
Then I read the article:
> We've analyzed why +80 startups have failed
If you've analyzed the startups then I really don't think that counts as "first hand".
Yep, that's what marketing is (having studied several subjects in it as part of my business degree back in the 90s).
Marketing is not just a synonym for advertising and promotions, etc.
Marketing is about finding/building/reaching a market for a product. It involves both identifying/creating a market, and identifying/creating a product for that market.
Advertising/promotions is just a small (though expensive) part of it, but it's the part that the world sees, so laypeople can easily think that's all it is.
This is mind blowing to me if this is true. I didn’t take many business classes so I wasn’t exposed to this definition. I do run a startup and as others have eluded about startup tech CEOs I’m pretty bad at marketing. But saying that, I still can’t wrap my head around how a marketing person can identify a market or create a product for that market. I basically created the company cause I saw a need. It’s my job as CEO to make sure our company is laser focused on solving that need. Solving this need is basically our overall focus but I understand you are probably talking about marketing creating new markets within that scope. Our process of identifying and creating a product and market is all about meeting with potential customers and trying to solve their needs which is a combination of sales engineers and sales and tech leads being consultants and figuring out how our technology can solve their problem. I don’t see where marketing fits into this such that they could go out and create a new market and new product for us..i always thought marketing’s job was to help us better reach those markets, but I had no idea they could identify new markets and create products for those markets. Maybe we see are using different terminology.
I've heard marketing used with this definition. It's actually really common - most marketing textbooks and classes will start from the premise that the goal of marketing is to identify a market and better serve it, and then work forward from it.
That said - I don't think marketing is a good goal for a startup, for the reasons enumerated here [1]. Basically, you can visualize the roles in a company like this:
And in a startup, you don't want anything to be abstract, because there's too much room for error just getting the concrete stuff right. Everything in the right-hand column belongs in a big company, with millions of users and thousands of employees, because they deal with emergent phenomena that happens when you're huge. If you're thinking in those terms you'll get crushed by the big companies that actually do have all those resources, so startups need to focus on small niches that can grow.
There was a time my title was CTO, and I'd say to the devs, there is only one function in a company. Sales. Every minute you spend writing tests, is sales. Because at the end there is only one business model. More money in then money out. I'll frame it in a different fashion, often the issue is that developers believe that they are alone with in creating value. And that all of these sales and marketing clowns are just.... What you may probably come to understand is that.... you are a marketing person now! and that this does not make you a lesser person. Yes, you can still code. But your primary function had changed. And over a longer run... it will require you to gain new competencies. So how does a marketing person fit? simply. They are you.
Event "product market fit" feels like a sugar-coated way to say that no one is interested in purchasing your product.
I don't have a startup, but if I did I would view every single step in the process of creating one, both technical and non-technical, as an opportunity to test the market viability of your product.
I'm honestly surprised at the abundance of non-product-viability issues that were cited as determining factors. Maybe this reality is hard to face?
You can't even say that a product failed because nobody wanted it. All you can say is that nobody you asked was willing to pay what you were asking. I've seen plenty of promising products die because the company wants to only offer it as a service with ridiculous licensing requirements presumably out of fear that some customer is going to build a billion dollar business out of it and not give them a cut of the gross.
It has to be really frustrating for startup CEOs to fold and then see a bad copy of their idea appear as free software and explode in popularity.
There are definitely execution failures, but in my experience spanning 20 years, I have seen far more failures that are really just failed experiments rather than execution failures. A half-decent team can get product-market fit; once you're there, attracting the people who won't fuck up the next phase is relatively easy. If you don't have product-market fit, no amount of genius will save your company.