Hacker News new | past | comments | ask | show | jobs | submit login
How a Large Customer Can Kill a Startup (howardlove.com)
197 points by imartin2k on April 9, 2017 | hide | past | favorite | 56 comments

Excellent article. And I want just to add more real world problems when dealing with large customers:

- they want to have multiple 'demos' with 5 or more attendees which end up being hours long

- you need to work hard to find out whether people you are talking to have any purchasing authority

- they want to have PoC deployed

- they want to have a kind of reviews and special contacts: security, special contracts, special NDAs, etc.

- they never pay invoices

- they are always want to tweak how the product works

In short, enterprise sales are more about relationship and negotiation - so you need to have team dedicated for doing only that (sales, sales engineers, etc.). It is not hard - but just you need to have team and process for doing that. For example, many established companies will have "paid PoC": You want to test our solution? Great. You have to pay first to test it.

Hit right on the nail with every point. I'll add one more: feature requests that are fishing expeditions on the customer's part, but are interpreted as show-stopping issues by sales/management.

And one more: byzantine network architectures and change management procedures that take six weeks and seventeen different signoffs to get a VM provisioned or an account created.

Always, always push back.

When I'm on a call and the customer/prospect* asks for something we don't have, I push to make it concrete. Here are some of my questions:

- How are you doing that now?

- Where is that in your roadmap? Is that required for the current phase?

- When do you plan to implement/use it? Is this 3, 6, 12, or 24 months out?

- Is this more or less important than X? Where X is another project or initiative, because there always is one.

- Is that a nice to have or a must have?

* Until they give you money, they're a prospect, not a customer.

When a big customer makes a feature request, call it a change request and tell them you'll make them an offer (they pay money up front) for it. Sure, if they never adopt the product the effort could be better spent elsewhere, but at least you'll get paid for their bullshit.

If it's expensive for you to do then you should always demand real money up front.

As someone who has 10 years of experience in EnSW, this is very f'n good advice. To all of the techies out there dealing with enterprise sales - close the sale, close the sale, close the sale. Worry about product later (and all of those are acceptable answers).

"feature requests that are fishing expeditions on the customer's part". This is a red flag. If your product market fit is strong, and you understand your value proposition, your prospects requests should fall somewhere on the roadmap. If they are too far out in left field, you may have other problems. The more your product solves a real problem, the more willing enterprises are to do things your way (take a patch every week, forgo some network architecture standards, etc). When selling, look for the signs that they are willing to play ball as an indication of commitment.

Those feature requests might be, say, integrations with other software they run (often one-off internal software) that they'd otherwise end up doing themselves. They want to see if "just asking nicely" will be enough to get that work done, rather than having to plan a whole project around it (even though it obviously would translate to you having to plan a whole project around it.)

If you want to watch that as a form of entertainment, I can recommend the "Billion Dollar Buyer" show (you can find it on yt).

i would just add that the proof of concept can cost upwards of six figures.

in other words, real companies that make real things charge their customers real money.

if there's no money, guess what, they're not a customer! move on. the business sociopath who you're interacting with at the other side of the table is more than happy to have you work for free for as long as you're willing.

I've worked for ten startups during my career. Even before the last couple, I had learned to recognize the "one big customer will bring us to the promised land" pattern as one of the surest indicators of failure. Those customers always end up sucking all the resources away from a more generally appealing product direction, and they never come through financially, so your development ends up months or years behind where it should have been with nothing to show for it. Some big companies deliberately take advantage of startups to do free work, with never any intent to buy the product itself. If you see this happening at your company, my advice is to start planning your next move.

My favorite brush with this phenomenon:

Startup I worked for is approached by a large multinational corporation for a demo project, something that if integrated with their logistics could bring us a ton of money and raise our exposure. The initial outlay from the multinational was only $50k but the size of the promised follow-on contract would really validate our out-sized Series A round that had been raised years before. About half of the engineers were pulled off of their normal work of gradually improving the product and the API to build this crash project. It was a really hard slog but the team pulled through and delivered something great, but it took around two months and the end product was very specifically tailored to that client and not generally useful for the rest of our customers.

Our team presented the final product, thinking we were about to kick the door down and be showered with a fat contract . . . as I understood it the rep from the multinational ended up not having any purchasing authority, he was an ambitious ladder-climber trying out a long shot with a little leftover budget he had. He had commissioned the whole project unsolicited from his managers and when he tried to skip up the chain with his genius plan he was shot down pretty hard and we never heard from him again. Meanwhile the normal work for the rest of our tiny but paying customers was stacking up and we had half the team demoralized and fatigued. Alas.

Responding to myself based on comments elsewhere. I've worked in enterprise storage, supercomputing, etc. In some of those markets it is indeed essential to have at least one big fish as a prospect, but that should as much as possible remain a sales-side phenomenon. Impact on engineering should be minimal. When it affects engineering too much, as another commenter said, you're no longer a product company. You've become a consultancy. If that fits with your business or exit plans, great. Enjoy your reward to the fullest. However, most of the time, you'd be betraying and failing people who wanted to join a product-oriented company and make a bigger difference.

Even large startups. Starbucks cost Square some $80+ million.

Can I get more info on this?

Been there, great indicator.

I think one of the puzzling things about start-ups is a lack of appreciation for basic business analysis models such as Porter's 5 Forces.

Having a large customer means handing over a lot of power to your buyer. Although they may empower you, they can also empower your competitor. Another way I like to describe this is Australia's contractor law, which states something along the lines of: if more than 80% of your income comes from one client, you are an employee and not an independent contractor.

Most start-ups only seem to care about competitive forces relating to industry rivalry, threat of new entrants, and threat of substitution. Start-ups are usually very aware of substitution threat as companies that claim to be "disruptive" are attempting to be the substitution threat to an incumbent.

However, many overlook the power of suppliers and power of buyers. A large client demonstrates the power of buyers. The situation between Netflix and Starz shows the power of suppliers, or Spotify and the music industry. Anytime someone is trying to build a platform with vendor lock-in, they are trying to build their power as a supplier.

I hope for start-ups that when valuing an investor, a key criteria is their ability to provide a business perspective that may not be immediately obvious to a technical or sales/marketing focused founder. It may also be beneficial be for more people read up on business books/practices as they embark on their entrepreneurial journey.

Very interesting analysis. Could you list business books that you like?

I very much agree but Cygnus was saved by two large customers and not sunk, but not really helped by one.

The big early customer that really saved us was Intel. The saving wasn't in the money (though we were bootstrapped at the time and so needed the money) but it was early days and we were not particularly competent at anything but writing code. Intel needed an outside party to make a i960 compiler so they basically came regularly down to my house (where the company was located at the time) to lecture us on how we should be getting things done.

NASA was another key early customer who wanted us to succeed to kept us between pretty narrow rails that we learned from.

And Sun...well they just gave us cash, were intermittently demanding, and very slow to pay but fortunately our income wasn't dependent on them so we came out OK.

This is also dangerous for consultants. I always ended up working for one well paying customer who used up all my time. When that customer goes away you have to pretty much start from scratch.

Most of my indie-or-small-team consultant peers have an anchor client that drives 90%+ of their revenue. And most of them make substantially better money than I do, as their billing utilization rates can be nearly perfect.

I'm admittedly slightly envious, as I have to work quite a bit harder for each dollar of revenue and my utilization rates are more in line with traditional professional services firms.

I do, however, take comfort knowing that my largest customer is ~8% of my annual revenue. I feel like I have a more sustainable and resilient business. And I think business development is something that needs to be happening all the time, not just when the threat of imminent starvation looms.

Another unexpected benefit of this arrangement is that I can walk away from "bad revenue". Since I'm not under pressure to grow or feed the beast of a substantial payroll, I've been able to start being more selective about the projects I take on. Once you've been in this business awhile, you can spot bear traps ahead of time, but that only matters if you're able to maneuver around them.

I think you have much better chance long term and you can probably scale up if you want to do so. With the anchor client you are stuck in a very comfortable place but you can't raise rates, you can't do different things or hire people.

And worse, even as a consultant, you have a sample size of 1 with which to base all of your decisions re: what customers want, what solutions are best, etc.

Which makes you even less appealing to other clients.

> The moment one customer pays you a lot more than any other customer, you’re no longer a product company, you’re a services/consulting company again


The dangers listed in the article are true, but it's important to remember that as you grow, you'll always have a "big one", one customer who is more valuable and demanding than everyone else. Once you finish digesting your first big one, and close a few more their size, a new, even bigger "big one" will come along and the process will repeat a few times.

In the early stage, you are looking for users of your product who all will tell you about the common usage patterns so that a product can be made. The previous startup I co-founded started selling to couple of large customers because the CEO thought that we needed to start at the top. I actually signed up one of these guys. However, the customer never actually told us the common usage patterns to help us develop the product rather they told us their usage pattern and hence it seemed like consulting rather than building products.

My experience with large customers in my previous startup was that they do all those things: they want demos, they ask for large volume discounts, they wants custom features and a custom license agreement and a 24x7 phone support.

My solution was simple: just refuse to do all those things. To my surprise vast majority of them went ahead and still bought my product (some posed an "offended" stance initially).

I guess for that to work you really have to have the best product for their needs. :-)

This article left a lot of questions open--what's 'large', what kind of product are you developing and how are you building the company? Depending on the answers the article's advice may be completely wrong.

Fer instance: when creating a product that depends customer design partners, small companies typically cannot afford to make the investment necessary to hold up their end. You'll lose the deal and burn a customer in the process.

p.s., Oracle and the CIA deal is a canonical example of 'one big customer' working out. [0]

[0] http://www.businessinsider.com/the-cia-made-larry-ellison-a-...

There is also a similar story for Siebel (Northwestern Mutual I think?). In enterprise software, buying from small companies is not the norm, so if some large organizations invest in your product, it can trigger the larger market to believe in you. I believe the trick is to make sure the "big" accounts you take on aren't ones that are going to take you too far off your normally planned roadmap. If you find yourself building one-off features, one for each customer, you are in for a world of hurt. If your big customers drive you to build your roadmap faster or more aggressively, that isn't always a bad thing.

For enterprise products the obvious 'large customers' to avoid are companies like Google and Facebook. They operate at such enormous scale that the problems you see there are unlikely to pop up in normal corporate use. If the article had mentioned these I would agree completely.

In my last company we had one particularly large customer that really put us on the map. It was grindingly hard to get things to work but fortunately they had a problem they really needed solved and just enough patience to let us solve the teething problems of new software. Once we had them as a reference it made a huge difference both in terms of product quality as well as validation of the company.

Yeah, but that's both the defense market (very unlike regular market) and the CIA (In-Q-Tel is unlike defense market). Quite an outlier that tells you nothing about the topic unless we're specifically trying to make that one, big customer be an organization that drops boatloads of money on unproven suppliers with a VC arm looking for them. In-Q-Tel is cool in that they've done this for a number of firms with great products. They're just really, really different from the norm.

This is basically the typical pattern for successful enterprise start-up exits. However, the only way I've seen this work out is if working for that big customer is instrumental in the start-up exiting quickly, and this requires some experience by management to understand if a big customer is a potential acquirer or someone just trying to get out of working with a bigger company at typical enterprise b2b transaction profiles (very expensive, very slow). If so, then the costs of that customer are better weathered by a bigger company anyway because bureaucracy and long sales cycles are not a specialty of any start-up in the world. But admittedly, over time this results in big companies taking on the technical and relationship debts of the companies they acquire and the larger company becomes slower and less agile.

Disagree. If managed correctly a large customer can help grow your startup in some great ways:

- cash to invest into growth, especially if you can derive a higher profit margin than average from the customer

- increased stability ("battle tested") in the eyes of other large companies

- pushes you to build a scalable and secure product earlier on

In my experience with Codevolve (where we prefer to work with multi billion dollar companies) I've found there is a balance between how much they need you vs how much they are willing to bend to your startup mold. If they want you enough, they'll concede certain points to you. Some important things to stick to your guns on are:

- included support (you want to provide as little as possible, and get paid for it as much as possible)

- non competes (you don't want them to get one, or at least not for very long)

- pricing (leave yourself room to grow the account, especially if you have a product that can be used across multiple departments, and you're just starting with one)

The best way I've found to do this so far has been to go "bottom-up": find a champion in the company who will actually use your product. They're much more likely to work with you than an exec.

There's a problem with the word "customer" here. Sure, having a large customer - i.e. somebody who has paid or is paying you - is fantastic in all the ways you mention. Unfortunately, what the OP was really talking about was prospects. People who dangle hopes or promises of some huge future gain, but with the condition that you have to expend a great deal of your own resources and use up your own runway to collect, are not so beneficial. They're the Nigerian email scam of business strategy.

"Greetings, X. I have many millions of dollars I'd like to spend, and I'm willing to give you a share. To facilitate this, you need only devote 50% of your company to me for a year. Please start working on features Y and Z immediately, so that I can make you rich."

Yeeeeaaaahhhh, not so much.

This is what I heard in Korea (someone with more insight, please correct me if it's nonsense): Doing a startup doesn't make sense because you'll end up as a contractor for one of the chaebols. Once that happens no other chaebols are going to touch you. So, in the end, you'll end up as an employee of the chaebol just without the stability and prestige that comes from working for a chaebol.

chaebol - noun - (in South Korea) a large family-owned business conglomerate.

Walmart is famous for really pushing this. To the point where they want to see your company books and decide what your margin gets to be.

I've worked with Walmart a few times now, and they want to invest in innovative start ups, however they are also risk adverse and they want to mitigate that. Perhaps they look at your books to see if you are going to be around next week/month/year and the side effect is insight into pricing. They definitely push you to deliver, however it's usually pretty pragmatic stuff.

I've seen issues with this a few times and had to bite my lip quite hard when the cost of the software we were providing was compared to the (low) number of Porsche cars they could have bought instead. No, a Porsche couldn't do the job they wanted and neither could anything else they could find for the money.

I'm just starting out trying to get a new product off the ground but finding I have an issue just getting one customer as it's something most suited to organisations with large numbers of employees. Either I get my own big one or I'd imagine the product gets a slow death.

Is it in a state such that you could sell it to a larger product or consulting firm serving the kinds of companies you were targeting? Either as a finished product ready for them to take over or as an acquihire?

I don't agree with this, my experience is the opposite, the sales effort required to sell to large and small customers is not that different, whilst obviously the reward from a large customer is much greater. Large customers are often more willing to take a risk with a small venodor, beacause they have much better systems and procedures on place to deal with such risk and most importantly they can better afford the cost, if it goes wrong.

> my experience is the opposite

I suggest that your experience is not universal, and might even be anomalous. Please, look at other comments on this story. A significant majority of the stories people are telling involve long lead times, expensive POCs, unpaid invoices, and so on. There's at least one mention of a "customer" who it turned out lacked the budgetary authority to make the purchase - something I've seen myself at two companies that fell into this trap. This is the same pattern I've seen every other time or place this topic has come up. Large companies are often more risk-averse because of internal politics and policies. Their systems and procedures are often the thing that make the sales process slower and more difficult. And they will use the carrot of that ever-in-the-future big sale to wring every last bit of free work out of a desperate startup. I think it's great that you've had a different experience, but only in the same way I think it's great if someone who walked into a den of lions got out alive. I wouldn't present it as the normal case.

The problem is not having large customers. Large customers are great, the biggest sign that you've actually made it. Large customers have deep pockets, it's like the business version of rich friends. They don't mind paying extra, even multiple times as much for you to go out of your way even a little. Rich people, and big companies, usually understand the value of your time and have the money to compensate you well.

The key is dealing with them properly, which this article implies is impossible. It may well be impossible for a start-up without employees who have dealt with these relationships before, but it's business as usual is the corporate world.

The most important thing to remember is the distinction between a product company and a consulting company. When dealing with large customers you need to draw the line and stay firm. As a start-up, you're almost certainly a product company, and you should not customize your core product for anyone under any circumstances. If they want something customized "that bad" offer to sell yourself to them. Walk away with a bunch of money to start your next adventure and a corporate friend that's probably willing to do you some favors.

That said, if they're going to throw money at you, might as well find a way to take it. This is when you build a stellar API, or even better, build your whole product on top of an API and offer blessed access... For a price. Your best bet here is to either partner with a consulting company or divide your company into two parts, one that makes the core product and the other that makes customizations.

The rest of my advice is a few general things for dealing with big companies that startup land might not know.

Number one is to make sure you're charging enough money. I worked at a rapidly growing consulting agency once and we had the problem of too much work. We simply could not hire and train fast enough to keep up with our sales pipeline. We could have done the naive thing and paused our sales pipeline. Instead, we charged more money. Twice as much per hour at first, then almost 3x. Any new work coming in had to be at a vastly increased rate or we simply didn't take it. Our reasoning? We were honest, we had too much work and the premium was to pay for overtime to get your project finished. We ended up vastly more profitable, we could offer more for the job positions, we could hire more experienced devs that didn't need training, and afford even more marketing. It's been less than 2 years and that single dev shop is now ten full time devs and still growing rapidly. A side benefit, we started working with bigger companies with deeper pockets because they were the only ones that could afford us. Not sure if there's an economic theory on this but it seems like the middle market is where all the competition is. If you're either at the bottom or the top it's a lot easier to make money. Remember this when making you're SaaS. Serve the bottom with your boilerplate product and the top by building stuff on the side for the big dogs. It's much harder to lose this way.

The other thing about big companies is that sales are more personal, not less. Large companies are more stratified, more organized. At a big company the decision of whether to buy your product likely rests with one person. Make sure this person likes you, a lot. When meeting with them send the guy he likes the most even if it's the janitor. At big companies the chance that the person buying your product is the one who will use it is minimal. While the quality of your offering still matters the most important thing is whether the guy buying it is going to enjoy dealing with you for the next ten years.

This ended up a long diatribe but I disagree in practice that having big customers is a bad thing at all, and argue that it's actually the exact opposite as long as you know how the handle them. I have worked at small companies with big clients and the name value alone (if they like you enough to allow you to use them as a reference) is enough to open most doors. Nobody questions your product or team if one of the largest most prestigious companies in the world is your client. It's the business version of getting a 4.0 at Harvard.

To add to my own comment after seeing a lot of others.... Big companies won't screw you if you have a solid contract. No contract, no work. Fight hard for NET30 and stop working the day the bill is late.

We never got fucked by a single client more than maybe 1k. Not once. Because we stopped working when we didn't get paid.

The economic term you are looking for is supply and demand.

That is, you were under high demand so your prices went up. Instead of, as you said, artificially restricting your own demand.

Large companies have big legal resources: if you mess up a contract with a big company, it could be game over for yours

I keep seeing this kind of info and meaning to curate it somewhere because I can never seem to find it again when I want to tell someone what a problem this is. I have finally figured out where I will curate it. This is the first article I shall so curate.

I love the closing.

Even better thing is having 2-3 big customers, where each one has his own priorities and you end up with multiple systems to be fit into one. So yes having 50 customers of which none has big enough vote to make changes in YOUR system is nice.

Yeah I agree, in theory a large corporation should have lots of money to pay your startup. However in reality being dependent on one big prospect to make or break your startup is a situation set up for failure.

Been there, done that. You should avoid dealing with large clients if your main target is small biz niche and vice versa.

At least with regards with the fintech industry, due to the myriad regulations one must pass to do most things, securing a large customer is sometimes the only way to survive, let alone enter the market. One of the main reasons we haven't been exposed to as many large startups in the space.

The best example of this is Bloomberg. The first 20 terminals went to Merrill Lynch (https://www.fastcompany.com/3051883/the-bloomberg-terminal). In fintech, I think both the regulatory hurdle you mention, in addition to solving semi-custom workflows can be a prerequisite for developing a useful product. Take Addepar for example. They have to integrate into the Bloomberg's of the world before they are useful. And, while specifically, a Bloomberg integration could be meaningless for the next customer who uses Capital IQ, neither are likely one-off integrations. Spending time customizing/connecting to other smaller/startup fintech services probably is likely disastrous.

I've heard a piece of general advice, that in general (if you are B2B) it is good to work with customers of a similar size as you! I think it makes sense.

It makes nonzero sense, but be careful of applying it too literally; very often, outfits of a similar size to you, either don't need you or don't have any money to spend.

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