Hacker News new | past | comments | ask | show | jobs | submit login
Before You Grow (ycombinator.com)
415 points by craigcannon on Nov 22, 2016 | hide | past | favorite | 164 comments

Building something users love is, of course, the challenge. 99% of startups are unable to do this, and some of those that can't still manage to raise money because someone involved went to Stanford or is otherwise well-connected, only to inevitably fail later on. Generally speaking, these days if you don't have rabid fans of your product almost immediately among some group of users - however small - the product as you have implemented it is destined for failure.

If you fail to attain at least some group of rabid fans right after launch, then you have to ask users and find out what changes you could make that would turn them into rabid fans - if anything. If you don't get some sense of agreement among users from the responses, odds are that you should move onto another project. Products that the masses love are very rare. Creating them is hard, and because it involves a tremendous amount of subjectivity and pure luck, it will likely take far more than one try. But you can only fail at it if you stop trying or let yourself get bogged down with attempting to make something work that users simply don't and will never care about.

VCs are optimizing for something other than love. They're optimizing for potential scale. You can build a product that customers love, but with a low ceiling for the market (like, say, building a really awesome surfboard). VCs want to see billion dollar market potential. So they may invest in something without real proof of customer love, or even customer need, just because the market potential is great.

That might be true at the seed stage, but no competent VC is going to keep funding you if you've launched and aren't able to demonstrate that some people love your product.

Of course, product/market fit is not a sufficient condition for success. You also need the market to be big enough.

On the other hand, the OP's assertion that money just flows to people who "look right" (my words, but true) without real regard for the product itself, applies largely to seed money only. A proper Series A round (or more) should require a strong product and customer feedback. So I think my point still holds. Seed VC is looking at market opportunity and founders who "look right", not a killer product that customers love, because the product may not exist or have a significant customer base yet.

"And if you’re just starting out, take the time to build a product your users love, no matter how long it takes. "

There's this weird balance between the lean "get it out as soon as you can, fail fast" and "make sure it's a good product before you ship".

What Sam meant was "take the time to build a product your users love, no matter how long it takes, before you focus on growth " (addition mine).

You should absolutely release your product to customers to gather feedback as early as possible, just don't fuel growth by spending before you're sure you've got the product 100% right and people love it.

What I think is missing in this is the matter of customer service. Get that right and everything follows, get that wrong and you are going to be slowed down by growth because far too much customer service time is spent along the way.

Imagine you sell a simple widget available in three sizes and two colours. A very simple product. Yet 10% of sales result in returns. If it takes you three days to respond to returns requests and then another hour processing a return weeks after the customer requested it then you are also going to have some damage done to your company brand on social media and customers that will not buy from you again.

It is not actually the widget that is the product, it is the whole sales process for that widget via the website that is the product. The customer service aspects of it need to be done really well before you grow. In the above example you can build out the forms for managing returns so that there is minimal effort needed, a scan and a mouse click in the warehouse, all communication with the customer handled totally automatically and procedures in place for the parts of the process that do need a human.

It can take weeks of work to get all of the customer service elements working properly, many of those aspects might seem to fail cost benefit analysis at the start, why spend ages making neat forms so your customer can return your product when you are really concerned with selling the product? Yet, if growth happens, this will be important. It could take an hour to process a returned widget when you take into account what happens in the warehouse, the office and in accounts. Or this task could be done in a leisurely five minutes if organised. The former is bad for the brand, the latter much better. You could get ten returned widgets a day, taking all day to do. This would deny you the time to build out those customer service forms needed for the five minute process to happen. You would need to hire an order manager person or a programmer to put the back end systems in place or else you are going to be stuck processing returns for unhappy customers, late, all day, every day. Maybe if you work hard they will not rate your widgets so badly.

So it is a classic step in time saves nine. But the product isn't just the widget, it is the distribution and customer service. Customers won't be able to love your product if they cannot get it or if you cannot sort their problems out with it.

Now managers with sales targets to hit could start advertising the widget before all of the customer service parts are in place. They may view too many returns as a nice problem to have and just think they can hire some interns if help is needed with that. Growth is gone for, widgets sell well but with the ten fold increase in sales is a ten fold increase in returns. Before long a whole returns department is established. This gains its own inertia and involves so many people that it becomes practically impossible to implement a simple form change to speed up customer returns.

The widget itself can be quite boring, it doesn't have to set the world on fire. The product is bigger than that, it is the brand, customer service and other things. This is what customers must love, and part of building a product they love is the dull customer service stuff. If you don't get those bits finished before you grow then you are just growing the amount of work for yourself.

Yeah the cognitive dissonance of startup advice can sometimes be jarring. In this case though I think you could say that fail fast refers to iterating on a product and changing it until people love it. Then the growth starts.

The poster child for not doing this in my mind is Adora Cheung. I remember watching her "growth at all costs" lecture from Startup School shortly after Homejoy shut down and thinking well, there you have it folks. Seems they suffered from exactly the leaky bucket problem Sam talks about here.

Two caveats: 1) I'm sure there are many other examples of leaky bucket companies, Homejoy just came to mind first 2) I still think Adora is brilliant and am really excited about her new cities project.

Part of the reason there is so much dissonance is the conflicting goals -- startup founders want THEIR product to succeed. VCs just need SOME products to succeed. So the advice that is tossed to cohorts of founders is designed to help some succeed, not to ensure the personal success of each individual founder.

Distinct actors and players have interests that are sometimes aligned or disjointed, but often contradictory. Though it may sound obvious, it clearly has not been fully internalized by plenty of folks. And when it has, few dare to draw the necessary conclusions that follow from it: always take everything you are told with a massive grain of salt. To be clear, I am not arguing that you should be a complete cynic, only that you ought to recognize that there can be a healthy dose of it. In particular when the stakes are high. Listen, but do not take everything at face value.

I am in an environment that has a particularly high density of genuinely enthusiastic, but naive, people; I am inclined to think that it is a mindset that's also predominant in other circles, besides University campuses.

Indeed, there seem to be quantities of people who are looking for silver-bullet secret sauces that will let them reach Steve Jobs status. Perhaps that's precisely because of the various personality cults that our industry subscribes to. I can't know for sure. And while I am sure there are some good anecdotes showing the contrary, it seems that, all in all, this is overall hurtful both to innovation and to the ones that try to accelerate it.

If you are going to be captain of a ship, you better not embrace any gust of wind that comes your way. Otherwise, you are more drifting away than going anywhere. And like Paris, you would rather be tossed by the waves than sink.

No, the rule still applies. Get it out fast so you can find out if your users love it. Just don't push it to a broader audience.

Iphone app companies do this one neat trick for this. They release in foreign markets (Canada, Australia, etc) and make sure their users love their apps before releasing in the US.

This is also one of the reasons New Zealand often gets products first - it's far away, super small, highly developed amd English speaking: http://www.economist.com/news/business/21651858-small-techno...

Time to release utes in the US!


Pity there is no company left making these in Australia.

I remember seeing supermarket checkout scanners in Vancouver BC and wondering what this space age (now there's a dated term) technology was and why it wasn't already in use in Silicon Valley in 1980. George Bush was similarly amazed ... in 1992. The UPC was developed in the US starting in 1966 starting from freight car identification.

Yes, late product release in the US even if the product was developed in the US is an old technique.

George H. Bush was "amazed" because he was being shown a newly developed model just being rolled out that scanned form multiple angles instead of just the bottom.

The way I understand it is:

(build -> deploy -> feedback -> build -> deploy -> feedback ..) until you reach love, when you can start focussing on growth.

The lean philosophy is to make sure your (build -> deploy -> feedback) loop is tight.

Get it out as soon as you can to gather feedback, possibly in closed beta. Wait before net promoter score is good enough before spending lots of money on marketing?

Reminds me of the Sean Ellis 40% survey: http://www.startup-marketing.com/using-survey-io

i don't think there is too much tension between "lean" and focusing on product quality before growth. the best way to improve the product is to get people using it and use their feedback to improve it. that's what lean is all about.

airbnb's 1000 days are mentioned in the article. it is not like they were just coding in their living room during that time. they had a product out, and they were flying to NY to talk to every possible host, improving the listings, and so on -- learning what made the product awesome. it was probably tempting at the time to work on getting more hosts on the platform, or to optimize the number of bookings, and other growth stuff. that's what Sam is cautioning against.

Build, Deploy, <loop> Feedback, Iterate </loop>, Hockey Stick, Investors, Monetize, IPO

I think it important to realize that everyone has opinion, and they are always different and you don't know who's right.

This is good advice but on a practical level isn't clearly compatible with other advice like "fix the plane while it's flying".

Also I was under the impression YCombinator focuses on helping companies grow and organizes a demo day 3 months later to help them raise money from VCs. The pg article even mentions growth targets during YC. If this is true then companies accepted to YC should already have a product people love and be ready to hack growth over the next 3 months. This clearly is not the case (eg Airbnb joined YC after about a year not 1000 days).

If YC is not only about growth but also about helping companies at much earlier stage still working on the product people love stage, that would seem to be a challenge. What do YC speakers talk about? Growth or Product? It's like putting kids learning Algebra in the same class as those ready for Calculus, not easy at all. Maybe there should be two types of YC sessions then? One for early startups still working on product, another YC for startups working on growth...

Every founder I have ever met wants their product to be loved, but unfortunately they need time and money to figure this out. Other issues: it's hard to know when to walk away pre-product and pitching VCs without growth doesn't seem to work. So founders raise money by showing early "growth" without PML (product/market love), then they go back to work on the product and when PML doesn't materialize in the first couple months they start to get nervous and switch resources back to growth hacking their mediocre product so they can be ready to raise money again and have another chance.

Obviously this is all very complicated, full of exceptions to rules. In reality product and growth have to be going on at once - but the timelines and resources and challenges to do product and growth at the same time should be part of the advice.

There is indeed a different YC program for early-stage startups working on product. It's called the YC Fellowship.


The YC Fellowship got axed. From what I understand, they're replacing it with an online course (similar to their startup videos/lectures) and it does not include any funding.

Oh that's sad - do you have link to an announcement? I wonder why..

I'm not affiliated with YC in any sense but trying to clarify some things:

"fix the plane while it's flying".

This is typically advice for startups that have found product-market-fit (i.e. have a product people love) but there are a million other things broken. In that case the advice is often to fix these things "while flying", i.e. while simultaneously growing sales and the company in general.

"organizes a demo day 3 months later to help them raise money from VCs"

I understand the purpose is primarily to help companies raise seed/angel rounds, not big "institutional" rounds. And just generally make connections for later.

"The pg article even mentions growth targets during YC."

Clearly this is not applicable to all companies. E.g. what's the "growth target" for a company like Rigetti who build a quantum computer?

"One for early startups still working on product, another YC for startups working on growth"

One of the primary (if not the most important) benefits of YC is the network. You get that regardless of whether you're in product or growth stage.

"pitching VCs without growth doesn't seem to work"

Depends on what you're pitching for. Seed and Series-A are two entirely different games.

"So founders raise money by showing early "growth" without PML (product/market love), then they go back to work..."

Good investors aren't fooled by that and will ask the right questions to quickly uncover any fake metrics. Besides that a seed round is typically raised on team, talent, idea and early traction, not growth.

>> Seed and Series-A are two entirely different games.

Are you actually in a position to offer advice? What you say seems outdated - seeds grew to multi-million figures such that "seed is the new series A" was a thing back in 2013 - firms now offer "pre-seed" funding and "second-seeds" seem increasingly common.

The assumption today seems to be that if you have no connections, you must self-fund / bootstrap / FFF your way to traction to even attract angel interest and for you, seed is for growth. However, if you have prior exits and connections then you play a different game and can probably raise a MM seed without even an idea of what your new business might be.

Are you actually in a position to offer advice?

Are you in a position to question that I am?

I've raised a seed round for my company 3 years ago and we are now going for Series A. I feel experienced enough to share some of my thoughts. Feel free to ignore it if you disagree.

Sorry I said that aggressively. I'm tired of low signal startup advice that does not address real-world complexity and it got the better of me.

What would have been fascinating, practical and not have raised my doubts would be if you compared the metrics and institutions you used to raise your seed compared to what you are using for your Series A.

Anyway, best of luck. Three years seems like a long time to live on a seed.

Thanks. We hustled to "Ramen profitability" so we don't actually need the money. We'd still very much like it though to speed some things up and to buy ourselves some ice cream to the Ramen every once in a while ;-)

What's the difference between "early traction" and "early 'growth'"?

You’ll have to rely on inorganic means like ads, marketing, or PR to maintain your growth, and this gets very hard to sustain

This is a point that I am always curious/skeptical about because growth will always become marginally harder. So ads, marketing and PR are always going to be necessary at some point.

Using his examples, facebook is creating airplanes to deliver facebook to people they can't reach otherwise - talk about high acquisition cost. AirBnB has a ton of TV [2] and Print ads.

Further, companies like abcmouse [3], dealdash [4] all blanket the airwaves with their ads and seem to be doing well.

So maybe some companies/products lend more to traditional ad spend than others, or maybe the idea that viral is the only way to grow and sustain is wrong.

[1] https://www.facebook.com/notes/mark-zuckerberg/the-technolog...

[2] https://www.ispot.tv/brands/oLp/airbnb

[3] https://www.ispot.tv/brands/d9N/abcmouse-com

[4] https://www.ispot.tv/ad/7Tc0/dealdash

The point is not to never do paid marketing, but to make sure that you have something in motion that doesn't rely 100% on paid marketing before you do. If people are talking about your product, paid marketing can help amplify the signal and the reach (facebook, AirBnB), but if you can get no one excited about it, if no one talks about it, then going all in with paid marketing is akin to shoving down people's throats something they don't want; it's unlikely to be sustainable even if you have large pockets.

CAC ~= CPC / conversion rate. If your conversion rate is 1%, then your CAC is 100 * whatever you're paying for a lead. If it's 50%, it's 2 * what you're paying. Many businesses that are uneconomical at 100 * their ad costs are economical at 2 * their ad costs, and even the ones that are economical are much more profitable.

The article is suggesting that you get your conversion & retention rates up into something reasonable before you blast out ad spending. You've got a lot more leverage here than on almost any other driver of your business's economics. It makes no sense to lose money on CAC - LTV and make it up in volume.

In my experience this goes hand and hand with getting your messaging right. If you have a product people love then you can work to articulate what they love about it and your marketing efforts become more focused and profitable. I agree that those channels can help develop your message, find early customers, etc. but that's different than investing in growth and trying to make your sales process repeatable; you're still learning at that stage. Investing serious money in marketing before you have the right product and the right messaging is the big mistake that I see a lot of startups run into.

I recommend using https://www.promoter.io/ to measure NPS. It's a brilliant starting point for getting useful feedback too. I've ended up using it for the conference and coworking community that I run as well as for a software product.

I would love to see a comprehensive list of all of these tools to measure NPS. Does one exist?

Sorry for the shameless plug, but just by coincidence, I have recently started building a simple NPS tool. Currently it's only a landing page, but I'd appreciate anyone interested to subscribe: https://nps.io

I have extensive experience applying the NPS to the hospitality industry, and I can tell you that it's not a marketing gimmick as I have read in some other comments. That's why I decided to roll my own service :-)

Serious question - why does anyone use a 3rd party service to request a 1/10 score from a customer? Kind of feels like it should be table stakes for a business. If they can't do that then wtf can they do?

The difficult design aspects seem to be when to ask and choosing what other supporting information to gather about who they are, where they came from and how they used your product/service - you can't just hook in an API.

You could try AlternativeTo: http://alternativeto.net/software/promoter-io/

Who finds actual value from Net Promoter Score?

It's a vanity metric, just like asking "would you pay $10 for this?" instead of getting someone to actually pay $10. https://en.wikipedia.org/wiki/Net_Promoter#Criticism_of_NPS

I don't think NPS is all that bad, but the one reservation I have is that sometimes a high NPS might be due to factors that might be unadvisable to emulate.

For instance, Apple has a very high NPS. A lot of Apple customers recommend Apple products to all their friends. Why might they do that? On reason might be that Apple products are high quality and their users have a good experience from using them. Another explanation is that there's a network effect and that Mac users want to be able to share programs and files and be able to get help when their computer breaks from their other Apple-owning friends.

A manufacturer of Windows laptops might see that their NPS score is lower than Apple and take that as a sign that their customers are more unhappy with their products than Apple's (which very well might be true), but it could also be the case that, in general, Windows laptop owners don't particularly care what brand their friends all buy because Windows laptops are more-or-less interchangeable. Company executives might conclude that adding more proprietary features to their products and creating a walled-garden ecosystem like Apple's is a good idea, when it probably isn't.

It's a vanity metric, until you realize that for your business, it predicts churn.

A lot of SaaS businesses have proved this for their customer base and found it's worth driving by. If it's not true for you, that's fine.

It's really useful to know if your customers enjoy the product with a 1 question survey. We turned a 0/10 customer into 10/10 by giving them extra love.

It's useful for us.

You're correct, the number by itself, doesn't provide a ton of value beyond a measurable KPI. For some companies however, that KPI is one of their core metrics. Some companies even base compensation around the score.

That said, the NPS methodology is much more than the overall score. The real value is in the individual scores, verbatim feedback and its proven ability to tie those results into both growth and churn. All combined (and when done correctly)it can help drive funding, determine PMF, guide your product roadmap, forecast revenue and drive measurable growth (just to name a few benefits).

Breaking it down to just a score is looking at NPS far to simplistically.

atrributing verbatim feedback to NPS is misleading though. getting feedback is just talking to customers, surveys and methods that have existed for years. saying its part of the NPS methodology is giving it credit for something it did not invent.

You're right, feedback isn't exclusive to NPS, but you're missing the point. It's the type of feedback that is the critical difference here.

Yes, you can ask any customer any question and get a response, but does that question/response elicit the same intelligent and predictive value of the NPS questions? Will randomly surveying your customers generate the same response rate as an NPS survey?

NPS is a proven methodology with predictive intelligence and high response rate, what you're suggesting is not. The choice is yours though.

"NPS is a proven methodology"

That's a blank statement that tries to be authoritative and sounds nice, but where's the proof?

Yep NPS is cargo-cult marketing at its worst. About the only thing going for it is so simple that even management can understand it.

Ditched it, and now only deal with a cut down version every year or so with clients.

CSAT has been far more illuminating. NPS often indicates success, while CSAT would actually show issues happening on the ground. It feels like NPS is idealistic and in the clouds, while CSAT is grounded and shows real feedback.

Plus, clients just don't like NPS either.

I think a good NPS survey asks why instead of just collecting a number and a great founder follows up to learn even more about why.

at that point its just a survey. why even prepend the buzzword "NPS" into it?

There is a clear difference between "just a survey" and an NPS survey. If you don't know the difference, I suggest you do some research.

What do you do if you make a B2B product that gives your customers such a big competitive advantage in the market that they won't talk about your product, actively hide that they use it, and refuse to let you tell others that they use it?

Sounds like a great sales pitch if you have the numbers to back it up. Walking into a meeting saying we saved a certain company so much money that they're scared to tell their competitors about it is pretty convincing if you can show them the numbers.

It is hard to back this up when the companies using it don't acknowledge they are using it.

Have you considered changing your pricing model so you get paid some % of their "success" however defined?

Edit: to explain my thinking - the idea here is it may actually be that if you're getting paid alongside them that you don't need to grow to other customers. You can get all the revenue you need from getting this one customer to dominate their market.

This is is actually what is currently happening in my market. My customers are increasing their volume each month while those that aren't using it are going down. The only problem is the market is quite fragmented so this approach only works if I have a customer in each market. I am working on solving this problem :)

Edit. I should add that I have offered a discount in return for being able to publicise their use. Unfortunately, everyone has declined the offer.

Then you want to bargain a discount, a faster delivery, a whatever, for being a public reference.

In other words, if they want something, and want it badly, instead of monetizing it with $$, monetize it with a case study.

Personal favor also. You go the extra mile for one user/infleluencer/decision maker in a customer company. Then ask to return the favor: "could you spend 15 minutes talking to this prospect of mine, it would help me a lot".

They can say no, find an excuse. But it will hurt them because they know you helped them unreasonably in the past, it was valuable to them, and you may be less willing to do so in the future.

Some will say yes. You just need one or two.

We also encountered some customers like that - true evangelists but refuse to partake in public endorsement for fear of losing competitive advantage. We even offered discounts, but to no avail.

On the contrary, we found the European and Australian customers to be completely different; competing companies in the same local markets seem to be quite chummy and will share such information freely.

Obviously not to be generalized. We have North American evangelists too.

I have tried to do this quite a bit, but they are all reluctant to be that person. Interestingly social proof seems to be unimportant in my industry, but it is a pain I can’t do case studies on my customers.

One of the more amusing things is different divisions of the conglomerate that owns my main competitor are among of my largest customers. I am always amazed how little internal communication there is within large companies.

Find ways to make your b2b product market itself by how it integrates (publishes content that is available on the web, REST APIs you can integrate with, etc). As your users derive and extract value from the product they will naturally advertise it for you.

This is the problem - my customers extract the value and want to keep all the value to themselves.

Yes, but by extracting value they are forced to integrate with the ecosystem at large. In this way, the ecosystem becomes aware of your product.

I am not sure what you are suggesting here. Could you be more specific?

Is this an actual case, or an imaginary what-if scenario?

Real product. I actually wrote about it over a year ago [1]. The number of customers now is twice as large and this is still a problem.

1. https://www.nucleics.com/why-do-only-5-of-the-93-dna-sequenc...

I use a B2B software from a startup and I would hesitate to tell anyone about it that I thought might go into the market. They're adding features based on my input, and I'm seriously considering developing the same software myself or outsourcing it so that it's not publicly available or the public version takes time to catch up to my own. (It's not about the costs.)

Raise your prices/get acquired?

I am not sure how raising prices will help solve the keep-it-to-myself-problem. Being acquired for megabucks does sound nice though :)

I recently came across an app lets you measure NPS directly on your website - https://www.metriculator.com

Seems really nice. I figured it might be useful for folks trying to run an NPS survey after reading Sam's recommendation.

I hate sites that have these.

IMO, if bothering the user at all, you should limit the number of options to two. E.g., after finishing a certain process like submitting an order, you can ask the user "Do you like our website? YES / NO". If NO, then you can take them to another form so they can complain with details. Having choices 1-10 is way too granular (we're not trying to get Yelp ratings). And popping up on the screen makes me want to leave the site.

On the positive side, I have had pleasure of using some well-designed interfaces, and the giving of positive feedback probably reinforced in my mind that I liked the site. I.e., it made me cognizant of the good design.

I guess it depends, but I find this less intrusive than an email.

Does a comprehensive list of all of these types of NPS-measuring tools exist?

A Quora post was the best I found so far. https://www.quora.com/What-are-the-best-online-tools-for-mea...

This is awesome. Way better than email.

Small nitpick: The blog is hard to read with so much bright orange distracting from the main content. I use a script to fix that on my browser. Thought the feedback might help with the new blog.

Orange, to Make Startups Great Again?

I try to let my own CSS (particularly fonts) on Safari trump the sites I read a lot, such as Hacker News. Safari's reader view is good too.

A lot of this has to do with experience and whether or not you're trying to fulfill expectations (VCs, etc). You really don't know when a product will become loved no matter how much you iterate and it's also impossible to know for sure how long you need to "slog" it out. Some products become loved far earlier than others. The best thing to do as a founder is to drive at your vision hard as quickly as possible, acquire users to test multiple theories and gather feedback, and iterate when there are signs. Even then, not all iterations or pivots are created equal. There are many different ways to grow a business (many of which veer from traditional silicon valley values) and it's up to the founder to stick to his/her own values and to what makes sense at the time. I don't think there's a single founder out there who wants to build just a mediocre product. This is why product-market fit should be the primary focus for any growing business.

i don't mean to be a dick but citing airbnb and facebook seems like bad taste. those companies are massive outliers with billion+ dollar valuations. nowhere near your average ycom startup.

i think i'd be smarter to cite maybe smaller companies that exited in the million/100,000 dollar range. seems more accurate.

i understand that ycom wants the people they're funding to focus on ideas that generate billions as that's ultimately advantageous for ycom, but, realistically, most of these companies will putter out or exit in the low millions, which is still cool as hell too!

doesn't seem to make much sense for anyone to constantly cite huge outliers like googs or fb or salesforce etc. when there's plenty of dope ass smaller ycom companies that probably did exactly what sam is talking about: build high quality, fun, exciting, engaging products.

FriendFeed focused on being loved more than growing users and had a $50MM exit.

It's a delicate balance, for sure. Focus on product until you have something that really works. But don't fall into the 'build it, and they will come' trap.

As a side note: the new UI for blog.ycombinator is really ...uncomfortable. 3 columns? 4 if you count the social share buttons jutting out on the left. Designed on a > 21" monitor :(

The list of tags doesn't need it's own column... each with it's own row. Just make it a tag cloud above or below the blog list, and give a little breathing room to the actual article.

The tags turn into a tag cloud at the bottom if your screen/window is small enough (I had to zoom out to get them to move to the right on my dinky x60). Smaller still, and everything (including the social share buttons) turns into one column.

  >> Airbnb slogged for 1000 days before discovering how to make their product loved.
Airbnb had a network effect product (marketplace of buyers and sellers). Growth accelerated once they had enough of both.

  >> It’s unclear exactly when Airbnb implemented what’s become their most famous growth hack, ... 
    Craigslist had one thing that Airbnb did not - a massive user base.

Everyone says to create a product that is loved (agreed), but how do you get constructive feedback in the early days?

I've experienced that many users are unwilling to engage in conversation, they test it and if they don't like it they leave and won't communicate why.

Suppose one creates a new Whatsapp-like product, how would one, in practice, go about gathering feedback?

The "unwilling to engage in conversation, if they don't like it they leave and won't communicate why" is your feedback.

Dirty secret that startup methodologies usually doesn't tell you: it's mostly trial and error. You observe people, you form a hypothesis about what they want, you put something in front of them, and you see if they want it. If they don't, change something. It's not unusual to require dozens to hundreds of iterations before you converge on something folks like.

This is also why founders that are building for themselves have such an advantage: then you do get some signal about exactly what you might need to change, and why you would want it. You've got to know (and accept) yourself pretty well to be able to listen to it, though.

Yes, I understand that it is a tacit feedback, but it is so binary that it is hard to infer something from it besides we need to change.

Iteration, it feels like a company could use the run rate iterating and never figure it out?

Most do. That's why most startups fail, after all.

The lesson to take from this is to get your iteration velocity up and your burn rate down. If you can, you also want to be extracting as much information as possible from the failures, so it's not entirely random, but even if you're brilliant and 100% paying attention there will still be a lot of them.

Most companies actually do burn through the cash trying to figure it out, and fail.

Early on, it's nothing more than an educated guess. Part of the problem here is that if you're actually creating something new, rather than a refinement of an existing product, you're in uncharted territory. Think of Henry Ford's comment about the customers wanting faster horses. That's not what they want. That's just the limits of their imaginations and experience. Great founders should be creating something the customers desperately want and need but couldn't even imagine or articulate. Henry Ford's customers just wanted to get from one place to another, faster than they could at the time, more reliably, and affordably. They wanted faster horses. He gave them mass-production cars that were affordable and easy to maintain.

If you're standing in front of them with a notebook in hand, it's hard for them to avoid saying something. So that's step one.

Step two would be to interpret nothingness as (negative) feedback that your product wasn't interesting enough for them to bother commenting on. Which probably means you're not solving a deep enough problem.

This problem is really hard for whatsapp-like products, since people already have their problems solved pretty well in that domain and so they are less likely to find a new product particularly interesting. If I were advising a startup who were making a whatsapp-like product then I would tell them to find a narrower niche where they could solve a deeper problem.

On a product I worked on we emailed every user that didn't engage with the product after sign up and requested feedback on how we could improve the service, what is lacking etc.. tried many different phrases but it was about 1 in 500 that replied and gave serious feedback.

Making current users even more happy is one way to go, they are more likely to provide feedback but it skews the perspective since we don't get any feedback on what the unhappy users were unhappy with.

Now, suppose it is a Whatsapp-style product, there must be a way to gather feedback even for them?

I suspect a lot of people that never replied would be people the signed up but never got around to trying the product properly or people that immediately saw that it wasn't ever going to solve their problem.

Would be interesting to correlate the emails with how engaged the user was after signup.

Personally as a user the only time I would try to engage the creators is if the product was very close to my use case but certain aspects didn't work the way I would need. This being for B2B where a potential customer is going to be more engaged, B2C the reply rate to those emails is likely always going to be tiny.

"""Startups are defined by growth, but growth isn’t step one in building a great company."""

I'd argue it is. Building something people love is all about growth. If they love it, they'll talk about and champion it. In fact measuring generic word of mouth growth is a decent way of figuring out if people love the product (my working hypothesis is that engagement is great but if they go beyond that...we're golden).

I'd say ignore all kinds of growth except word of mouth growth. I don't care about user acquisition schemes...as soon as the word of mouth growth flattens the alarm bells must go off and all other measures are probably just life extending measures.

I'd also say ignore "growth hacking" especially of the "I'll trick people to like my product" variety. Genuine product improvement is the best growth hacking there is.

Another alternative: don't build something huge, don't get funding from VCs, build a comfortable business for yourself instead of a gimmicky trend that will be irrelevant in a few years. I feel like this is really underrated these days, if you spend the time truly earning it, not only is it more rewarding but the company may last more than 3-4 years.

Build a small focused product to generate some revenue, then hopefully that will give you time and financial freedom to build more complex products down the road.

Is there any data on the success rate of companies that aim small? A lot of the allure of the startup is a chance to strike it big. If one isn't doing that anyways, might make more sense to join a big company since they offer a considerably higher expected return on your work.

(Also if one can get funding, it lets one get paid an actual salary more quickly.)

I have no clue, it seems uncommon in software, but personally I'd choose working 3 hours a day over 200k/yr any day.

IMO having a small team constraint is powerful too, you really have to optimize things otherwise you're screwed. Basically the opposite of a startup, where you'd toss money at the problem while you try and attract more people, making the product increasingly brittle.

working 3 hours a day over 200k/yr any day

A business that you own and which makes you any money on 3 hours a day may be even more rare than a unicorn. In most cases you'll be spending those 3 hours per day on administrative stuff alone.

That's definitely not true. There are only a few dozen unicorns, and hundreds of self-sustained businesses that "print" small amounts of money. If you join any large city's small business CEO groups (in NYC, SF, Toronto, London, etc.) you'll find a lot of one-person, bootstrapped companies that just run themselves.

It's not easy, but there are dozens or hundreds in many large cities.

Success as a small company often comes from passion and expertise. Passion wants more than 3 hours a week.

And if it makes money in 3 hours, chances are it makes more in 30 hours.

Three hours per day, not week :p. It's very doable, work smart not hard as they say!

> don't get funding from clueless VCs

Usually if people put money down and there is an entire industry around it, the collective mindset of that industry is close to optimal. They might be wrong on a couple of important issues. They also might be very wrong, but they're still close to optimal by definition because nobody's doing it better.

TL;DR: calling entire world-class industries clueless implies naiveté.

Removed my "clueless" generalization. It's anecdotal, but most that I've encountered (relating to SaaS products) don't have much of a technical background at all, just throwing money at buzzwords, meanwhile they think critical services such as logging aren't worth going after, just because it's not trendy. I'm sure it varies from industry to industry.

Hey TJ. It's easy to feel that way coming from a deeply technical background, but trust me, money isn't "thrown" anywhere. It's a high-risk high-reward game where one company (esp. at seed stage) will end up returning the majority of your fund, and every company has to potentially be that company, so outlandish buzzwords that might be the next big thing more often draw capital. Investors are people, paid to move money, with targets to hit and limited partners to provide returns for.

In the same way you or I wouldn't accept a PR that doesn't align with the larger vision of a framework or library we've written, VCs don't want to fund a company if the leaders of the company aren't empathetic with the VCs goals. It's about aligning interests. The goal is really to align technology and people - create something technically amazing, with huge potential, and align yourself and your vision with those who have the capital to fund you.

You don't have to do the dance if you don't want to, but I'd be hesitant to be overly dismissive of it.

Logging is indeed a critical part of many processes, but it's very hard to do it in a way that pleases everyone. Most loggers don't even support renewable inputs like bamboo.

Before you think this is a mere joke, consider that normal people don't even know what computer logging is, much less want to become a user of a logging service. The addressable eyeballs are few, and it's a market chock full of free products already. Plus the most serious consumers always end up with a bespoke proprietary solution.

Serious money is rarely thrown at anyone - no matter how buzzy their language - without some sign of a defensible offering that has product / market fit in a sector with healthy prospects + a team that can clearly execute against the opportunity.

While it's easy to find sensational stories to the contrary, bear in mind that sensationalism is a cheap path to sales, meaning not every narrator is reliable. Indeed, if you start raising money yourself you'll quickly discover how unreliable many "sources" can be.

In finance this is called efficient market hypothesis. And as I understand, generally considered false.

From Efficient Market Hypothesis: Is The Stock Market Efficient? | Investopedia http://www.investopedia.com/articles/basics/04/022004.asp#ix... "Conclusion It's safe to say the market is not going to achieve perfect efficiency anytime soon. For greater efficiency to occur, the following criteria must be met: (1) universal access to high-speed and advanced systems of pricing analysis, (2) a universally accepted analysis system of pricing stocks, (3) an absolute absence of human emotion in investment decision-making, (4) the willingness of all investors to accept that their returns or losses will be exactly identical to all other market participants. It is hard to imagine even one of these criteria of market efficiency ever being met."

I wouldn't say it's generally considered false, considering that the company founded on this assumption - Vanguard - is now the largest fund in the U.S. and second largest asset manager in the world.

Markets are interesting, though, in that the more people believe they are efficient, the more inefficiencies creep in, and the more people believe they are inefficient, the more they become efficient. When everyone assumes everyone else has found all the opportunities, nobody bothers to look, and so there are lots of undiscovered opportunities to cash in on. When everyone assumes they can get rich finding opportunities, they bid up prices to very close to their "true" value, and the profit opportunities disappear. It's a self-correcting system.

The question of how widely believed the EMH is itself suffers from selection bias. People who believe markets are efficient don't enter the financial industry, they give their money to Vanguard to manage. That means that everyone in the financial industry, by definition, believes that they can make better-than-average returns, otherwise they wouldn't be in the industry. And then by returns, most of them are wrong, but that also doesn't matter because the few who are right are the ones who survive to play the next round.

While true, assuming the collective mindset of an industry runs close to optimal just because its there is also naive. VCs themselves bet on industries being un-optimised!

Plus in evaluating an industry you would need to see it in context. So returns for investors VS similar asset classes. Or in this case how VC funding compares to other funding methods for founders under relevant circumstances - and looking at relevant metrics (pace of growth, sanity etc).

> VCs themselves bet on industries being un-optimised!

And they're usually wrong.

Not affiliated with this site in any way, but Indiehackers shares many examples of people doing exactly this: https://www.indiehackers.com/businesses

Indie Hackers is a great site! I have my first little product on there (http://apex.sh/ping/), making just over 3k/m now (50% each month FWIW), sure it's not millions but now I have time/freedom to work on the next thing.

Plus the freedom is worth thousands alone in my opinion, if I have the choice I'll never go back to startups. It seems like startups are the default these days, and so many of them would probably be just fine as regular small-medium businesses. I have to cram the idea in people's heads haha, you can make money in software and not be a slave working 14hr days.

You can also build a successful startup not being a slave working 14hr days. Both small businesses and big businesses can be either well run or poorly run. It's mostly a function of building the right culture, having the right systems in place, and having enough resources to not burn your team out.

If an idea works but you don't scale it to some sort of monopoly using your tech advantage or lock in systems a competitor with VC money will eventually outprice you out of the market, get an unbreakable hold on the business whole then convert some aspect of it to premium to monetize the now huge user base.

Depends on the product I suppose, if it's niche enough you may not have big competitors at all. I choose to entire a SaaS market that is already monopolized but I'm still paying my bills. That's definitely something to watch out for though, I think it's best to have many smaller products for redundancy, at least that's my plan.

This can work and be a great, profitable business. But a lot of these kinds of markets are winner takes all. Which is why companies compete with VC money to be that winner. You don't have great chances of doing that without VC capital if you have VC funded competitors.

But your objective may not be go big or go broke - and you may well end up personally happier for being more balanced in your life. That's perfectly fine in my opinion, and very underrated here on HN.

This is not necessarily true. For example, your market may be too small / niche to attract venture backed companies but large enough to sustain your lifestyle business.

Edit: Foiled by my delay settings! Damn.

100% agree with this.

IMO there is to much focus on becoming high growth vs simply creating revenue. Of course creating high growth companies is great thing to do - But so is creating a company that solves a small niche problem and creates sustainable revenue.

Another alternative: use this compensation model and you may not need VC


Have you actually implemented this idea in contracts? It sounds cool in theory but a PITA/unrealistic in practice.

Yes we pretty much did and why?

Well to name a few reasons ... seems infeasible/inaccurate to measure and attribute the various changes in the product to individual contributors or changes. (eg. what if multiple changes are occurring simultaneously, what about seasonality, what if it's just a random surge, etc.) Further, seems impossible to verify or trust metrics for the external developer. If the company is behaving unethically and misreporting/shortchanging, good luck suing them. Furthermore, from the company's perspective they're opening themselves up to massive legal risk by making this as a contract for the reason above - developer might try to sue them claiming gains which may or may not rightly be attributable to that developer's effort.

I think this is a situation where be a just ruler and exercise judgment in paying employees in proportion to their contribution, makes a lot more sense than trying to codify an agreement.

Do people have thoughts or pointers for building something users love when your product depends on there being enough users using the product? Dating apps, platforms and such.

When a single user gets no utility from your product if there aren't enough already using it, what does one do? You kind of have to spend some money marketing to get the initial word out. This is something I've been struggling with. We've been doing a lot of writing content and there's some word of mouth happening. We spend small amounts on FB too. But we're seeing a lot of users drop off with the reason "there just aren't enough other people using this".

You should look into improving your Single User Utility.

Here are some links to get started:




There are also countless examples of successful two-sided marketplaces who started out by "nurturing" (aka faking) one side of the market until they reached their critical mass.

Foursquare solved the issue with gamification (badges).

For your dating app example, you could add a non-geographically bounded ranking feature, similar to "hot or not."

Basically, creating any sense of accomplishment within the platform will make it more difficult for people to abandon, giving you time to reach a scale where the core product value can be realized.

Hard to give pointers without more details of what the product is.

There's all kinds of strategies, like starting hyper local (Facebook started just on Harvard's campus) or other startups have used waiting lists before launching to get a minimum threshold of users before opening the experience.

Be careful of "there just aren't enough other people using this" as it directly relates to Sama's advice. Again it depends on the type of product you are building, but when there's a product users love or see they will love even more if there were more users, they don't drop it so easily rather they go out and recruit more users.

What could you change to get these users to love the product so much they instead help it grow and recruit other users instead of just dropping it? I don't mean give them $20 (though you can) I mean how could you make the product so great they want to tell others to use it too?

1) New every time - I don't think there can ever by a free and repeatable way to do it because its so valuable that any method that gets discovered quickly gets saturated or gated to extract a % of the value

2) Bootstrap - grow from a small X that generates value for individuals/small-groups and slowly mutate into a Y that delivers value primarily from the network as your user-base grows.

3) High cost blanket advertising or PR - get an inrush of like minded people at the same time. I can't actually think of a success story that achieved this, plenty of failures though.

4) Better than 3) is to use capital to buy users and user activity such that the value stays in the network e.g. paypal giving away free $$ to every sign-up that they then paid to a paypal merchant.

Didn't Tinder start off by seeding your experience with friends of friends on facebook?

I believe Tinder did a lot of going to campuses and throwing parties, that sort of thing.

I really wish we could see some metrics on products people 'love'.

I'm working on a product that is seeing great growth (free atm), and what appears to me to be good customer engagement.

When I break things, users email and let us know, time on site is good, etc. etc. At what point do you go, yup, this is something people love.

Of course, I'm still making product improvements and figuring out what people will pay for.

I'm curious about how this maps to enterprise products? Can you create an enterprise-targeted product that users truly love?

Any examples of this?

The enterprise sales process isn't what it used to be. Before, you had to get an audience with the CIO (or whoever controlled the department budget) to have a chance.

The shift has been that employees and individual teams now have a lot of say in what tools they use -- sometimes without the higher-ups knowing about it until it's seen a good amount of adoption amongst internal teams. SaaS with instant access and 0 deploy times makes this possible.

Another shift is that employees will now carry their championed products with them from one company to another. In an area like Silicon Valley where job hopping is common, the social/word-of-mouth effects on enterprise software is significant.

Slack has already been mentioned and is a prime example of both.

With our own "enterprise" product, we're seeing the same -- adoption from teams and employees that eventually can become something more. But, this can only happen if you're resonating with some group of users that love your product.

Yes: https://www.Odoo.com

But we did not follow Sam's path. 1/ we made resellers and developers happy (being open source played an important role early on) 2/ we grew 3/ we made end users very happy 4/ we grew much faster

The reason we took this part is that it takes a lot of time to both build the enterprise features, and make it great. We built the scope first and made it great after, which is not an usual path.

Disclaimer: I am the founder.


How do you grow your reseller network?

I would presume that reseller's interests don't always align directly with end-user interests?

JIRA would be my example of it. Of course, it seems old and clunky now, but when it first came out, it was a revelation in a world that had been stuck with the likes of Remedy.

Salesforce is another example. Again, it seems dated now, but when it first hit the scene, it was seismic.

It's actually pretty straightforward to make a loveable product for the enterprise, if only because the pain can be so great. The problem is wading through the sales process.

Slack when it first came out. It spread via word of mouth.

My takeaway: Build a product that people will love (don't forget about yourself too!). If not, fix it immediately before it's too late/difficult.

Easier said than done :-)

But, I have a better advice: Do not spend money and time promoting something that sucks or it is not useful.

that's actually why best startups are those who make products for themselves, so they tested on themselves first; they don't really need public launch date; because the product is already being launched and getting instant feedback during development

examples are airbnb and dropbox

I have an even harsher take.

In the old days if you didn't have a product that "sold itself" (i.e. loved) then you'd just make it cheaper than the competition.

These days software products are by and large free (your users pay in time), so there is no obvious competitive strategy on similar products. If you don't have evangelical user base that loves your product and spreads it or at least a well defined niche use case, you've literally got nothing.

There is no room for "Facebook but cheaper". Heck there isn't even room for "Facebook but marginally better" anymore.

There's plenty of room for other Facebooks. In fact, there are major competitors to the likes of Facebook & Linkedin & Google.

They just happen to be national companies (russia, china, some european).

If anything. The thought that they have no competitors/alternatives is just a display that SV totally ignore the local non-english competitors (no matter how engrained they are in their market) and generally disregards anything that is under 1 billion unique users.

Actually I think these mostly succeed against Facebook where governments create an artificial environment by blocking Facebook.

Perhaps the most successful of the social networks that competed against Facebook on even terms was Orkut, and it's gone now.

Markets with strong network effects, like social networks, tend to be winner takes all. You'll never upset the market leader by making a marginally better product. If it's not 10 times better AND can't be copied by the leader you've got no chance really.

If Facebook is replaced it will be by people shifting to a different paradigm, not to an incrementally modified version of the same thing.

This is only for ad base products.

Tesla has far under a billion users and SV loves it.

Many b2b startups have a few hundred to thousand users.

If you want to sell Ads and be free, you need billions. If you have an actual product you charge money for, you need far less.

What if you want to sell users data? Isn't that what most free web and mobile apps are really selling?

I have a lot of great software ideas (great in that I wish they existed and want to use them), but no clue which could actually be monetized effectively. Billions of users seems like a tall order, and not a sensible thing to pursue.

Easiest is b2b. Are any of your ideas something a business would pay to use? Start with a product that cost say $50-$500 per month.

Much easier to sell 100 copies of something than get 100 million people to use your software.

Silicon Valley (or at least the large corps) is very aware of the regionalization of software adoption. Its simply that how success is defined for them is by, as you said, large numbers that often require something larger than small regions.

One of the difficulties of SV is that several people have made the "Facebook for Brazil" or the "X for Y region" but over long term iteration, the regionalized product gets overtaken by a supra-regional product's momentum.

Google had Orkut which was famously very popular in Brazil and India. Network effects couldn't sustain them when Facebook successfully went global and mobile.

I think Facebook acknowledges that they have competitors in china such as weibo,renren and wechat. That's the reason Mark Zuckerberg visits China so frequently to persuade chinese govt in allowing facebook access there to gain that enormous market share.

An increasingly popular strategy is "X but paid". People want to pay for software/products that is a real business, won't get acquihired, and will stay around for as long as they're users.

It's a competitive advantage to charge users. Because when user == customer, your incentives are aligned. People know that.

TL;DR - Two good ways to measure if you're making something people love: 1. Your users actively tell their friends about your product. 2. NPS surveys.

In the beginning it seems like the best way to measure if people are telling their friends is by not measuring at all. Just talk to users. You'll find out if people are about to tell their friends or how new users found you (an existing user sent them your way).

I'm always weary of surveys since they can be a dangerous mix of qualitative and quantitative. NPS seems solid since there's no denying a number. Sometimes I see NPS surveys that only ask for the rating (1-10) and don't ask a follow-up question ("Why did you choose 5?"). This is such a missed opportunity since the "why?" text can contain feedback gold. There's also no substitute for following up after the NPS survey to do more in-depth qualitative feedback gathering through user interviews.

I'm curious to hear more about what Sam thinks about measuring DAU/WAU/etc in the beginning.

Comment withdrawn. I've withdrawn this comment, but per site policy / encouragement I'm leaving my original comment below the lines.



I liked your comment but you won't believe how I felt about one part:

>Iphone app companies do this one neat trick for this.

I'm sorry, I just found it very distracting... like you use clickbait title interspersed in your writing. I tried to replicate the effect for you at the top of this comment. There's nothing wrong with what you said, it's just kind of like the format of clickbait titles, you know? Thank

EDIT: got voted to -1 but I'm keeping this - it's genuine feedback for someone. the phrasing "do this one neat trick for this" doesn't come from normal writing or conversation and I find it distracting in an otherwise nice and informative comment.

We detached this subthread from https://news.ycombinator.com/item?id=13016485 and marked it off-topic.

They're obviously parodying the "one neat trick" trope.

I don't think so because they're not talking about a marketing gimmick, but rather talking about an actual business strategy - I just found it distracting. They could have at least put it in quotation marks if they really meant to include it in their writing for some reason. I've withdrawn the criticism.

I don't know if it's a "gimmick", but it seems precisely "marketing".

I think you didn't get the point blazespin wrote: they said (or I read it as them saying) that since there's no way to get a "test market" in the U.S., rather than blow their "one shot" on the launch version which they can't iterate on before the whole U.S. market sees it, they use a proxy for a U.S. test market: they use a smaller foreign market. They don't care about "burning" those users or losing their "one chance" to launch, since those markets are small or at any rate a smaller launch. They then take their learnings, incorporate them and launch in the U.S. with a first version that is markedly better.

so it's product development. the point isn't about marketing at all. (Except in a broad sense.) it's a product development trick, like a closed beta. I don't think you would call a closed beta "marketing" exactly.

at least, this is what I read the comment as.

That was the point I got from it. My point was that it sounds a lot like what's done by the people I know who work in marketing at big companies. Marketing is not just (or even mostly) advertising.

all right :)

I don't know what "normal writing" is, but I hope it's not a virtue to be boring.

Have an upvote, that cliche was very unpleasant.

Saw the title, assumed it was going to be about weed.

Wake me up, before you grow grow.

(Sorry, could not resist)

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact