Hacker News Reader

Comment by pc | Hacker News Reader

We're excited to hire Mike in part because AWS has been one of the very best examples ever of simultaneously selling to the largest companies in the world while not giving up on a strong self-serve product. So many B2B companies end up with homepages for which the only CTAs are annoying variants of "read case study", "download whitepaper", and "request a call". All stock photography; no product imagery.

Stripe's growth thus far has been fueled by individual developers and founders signing up and growing with us to billions of revenue. As we scale, the "impatient developer" will remain the focus of our product development.

Over time, however, we've been increasingly finding that large organizations (like Zoom/Atlassian/Maersk) want to adopt Stripe -- Stripe now has a lot of functionality that they can't get elsewhere. So, we're very eager to continue to grow this side of our business with Mike.

View on HN

Congrats, Patrick. The VMware + AWS experience is certainly a big plus for a company like Stripe.

I would dare to disagree on, or at least to challenge, the "simultaneously" that you mention, though - in my experience (ex AWS, ex VMware) these two are completely different sales motions, made possible simply by the product itself and how it's discovered and consumed... Despite (yes, despite) what enterprise sales wanted to do. In a way, it's as if AWS self-serve were one entity, and AWS enterprise sales were another one.

In many cases, sales teams were completely not aware of certain things happening, or certain accounts becoming huge overnight, or certain developer behaviors. Some salespeople were handsomely rewarded for having been randomly selected as the sales rep for accounts like, say, Dropbox or AirBnB.

It's hard to describe properly, I know, but it's not as harmonious as one might say at first glance. And I think that the sales team still hasn't a good idea of the other side of the coin.

In a way, AWS' merit is to not have let Enterprise Sales screw up the self-serve part, and I think that's mainly because AWS was, and still is, so much more metric-driven than any other business I'm aware of. Enterprise Sales happened gradually, without interrupting a very addictive self-serve machine.

Anyway, this could be a long conversation with lots of platitudes, so I'll stop here and try to think of better ways to explain what I have in mind.

Yes, the world certainly hasn't yet nailed this playbook. AWS has, I would contend, objectively done an incredible job in this synthesis. (How many other businesses have grown to $40B+ of revenue with self-serve product adoption front-and-center?) But I'm sure Mike (and anyone else at AWS) would be among the first to agree that they also have things to learn and improve.

Yes. I once calculated that AWS is the fastest ever b2b in history. Only marginally slower than FB, which is a consumer business (and the fastest ever).

One more thing I'd like to mention in this context: ask 100 people why AWS has been so successful.

Think of your answer.

My answer (6 years at AWS; met with ~5,000 AWS customers, so big data set): lowest barrier of entry (swipe credit card, EC2 is up in 2-3 minutes), and lowest barrier of pausing the service (and the billing).

Most people would say any other thing, but not this one. In fact, EC2, S2, EBS were initially quite bad, in terms of performance or price/quality, compared to pretty much anything out there. But you needed 2 minutes to spin up an EC2 instance, vs 2-3 weeks anywhere else.

The other thing is trust. They have (yet) to screw customers on pricing that I've seen. Has amazon ever raised prices on an existing service? I'm not saying the prices are any good, but they don't seem to go up.

So you can spin up in a few seconds, scale, stop and even invest in the platform without worrying too much. Also - they don't seem to discontinue services that quickly, again, doesn't feel like you will be screwed.

I just had a call TODAY with a vendor, the price could be as low as $30/unit or $120/unit - that's a near 4x delta from sure to hell no for our use case. Sales guy needs to go talk to their manager to see if they can get us the lower pricing. And who knows if next year (it's annual) we'll again get their "manager" to sign off on the lower pricing.

Something about enterprise sales folks thinking is very short term. No price transparency - maybe we can qualify and squeeze them for more per unit. Or maybe get them in low and squeeze them on renewal once they've invested. The overhead of dealing with them is so high too. Show me actual product screens (not talking heads and bullet points). By the time you get to the demo you have wasted a week of your life.

So big thumbs up to AWS here. You can see baseline pricing up to huge traffic. You can see savings plans discounts up to large amounts etc.

AWS has clear upfront pricing, but they're a large enterprise company so if you're a large customer, and you're not getting a discount, you're definitely getting screwed. No one talks about that because it's all covered by NDAs covering backroom deals, but you can be sure Netflix isn't paying list pricing for EC2 instances.

Having to convert between Reserved Instance families to save money, while tricking people into thinking that this is good for them, is a masterstroke of AWS' sales department. Wouldn't you rather be working on the product rather than optimizing RI counts?

I've never felt tricked by RI. Is there some value to a cloud provider if I commit to three years of an instance type that might night have much new adoption after 1 or 2 years? I suspect so. They seemed to price based on that. What's the "trick"? Standard a1.medium instance for 3 years runs $7/month. Not unreasonable

If you don't want to do that use savings plans.

If you like talking to sales people and negotiating discounts do that. But they haven't let those sales people shut down / mess up self serve, including options for discounts.

You can get pretty far with amazon's tiered pricing and savings plans / spot etc discounts.

At the time EC2 became popular, self-service rental of virtual machines was already a well-established business, but it was called "VPS" instead of "cloud."

Uploading source code to your provider and letting them figure out how to serve HTTP traffic from it was also a dominant pattern, but was called "shared PHP hosting" instead of "serverless."

I'm not even old. But I guess most people my age probably did not have HostGator accounts when they were 11...

You have me beat, I don’t think I got my HostGator account until I was 13. Also not old.

My girlfriend got a traditional CPanel hosting setup a couple of months ago. Being so caught up in my tech bubble, I kind of forgot such quaint things existed.

But the more I think about it, the more I wonder why I’m bothering to write a service with JavaScript, fiddling with webpack, or deploying to Digital Ocean or Heroku or some other overpriced tech-hype VPS service, when I could instead simply drag and drop a .php file into FileZilla and get on with my day?

AZ recently wrote a well-reasoned piece about the need for many companies to understand, and to be great at both growth and sales to open new market segments: https://a16z.com/2020/07/29/growthsales-the-new-era-of-enter...

As you acknowledge, there are few companies that have made the transition from growth to growth+sales in the way AWS has. If Mike can help Stripe chart that course and avoid the shoals along the way, that's hugely positive for customers, employees and owners of Stripe.

I learned about AWS and Stripe at around the same time, but while AWS's dashboard has become truly overwhelming, Stripe has until recently been very focused on your core product.

I always imagined that flexibility was a key part of selling to enterprise while competing on something other than price. You also see this in other verticals: Salesforce in sales, Netsuite in finance, Twilio in comms, etc.

Stripe has already launched a few new products over the last few years. Is this hire a further indication that Stripe is shifting toward being the provider for everything payments? Is the Stripe dashboard going to be the next to fall to the curse of having a thousand buttons?

(I don't think it's a bad thing, it's been making everything a lot easier for us. But wow it's hard to get students up to speed on AWS compared to when I started using it :-))

I think that the infrastructure market naturally breeds a lot of products for pretty specific/narrow use cases, whereas our strategy is more about marrying financial capabilities with very broad/flexible APIs. Given that, we've always valued very cohesive and composable functionality between the different facets of Stripe, having a straightforward and tightly integrated Dashboard, and so on. (That's no criticism of AWS -- we haven't launched 100+ different products and we aren't going to have a product suite as extensive as theirs any time soon.)

I do think that software complexity lifecycle management (and how that pertains to organizational considerations) is one of the tough problems in our industry. AWS has done an amazing job in terms of their pace of development and innovation, but, yes, it's also become a pretty complicated suite. I think we all currently face some kind of "fast development", "cohesive integration", "broad functionality" trilemma and how best to surmount it seems a big open question to me.

Thanks for the reply (and the fun new word trilemma). Accessibility for new developers has been something that's been on my mind a lot recently. Glad that it's something you're mindful of!

I think the answer to the trilemma is simply more people in the ecosystem. As products grow to become multi-billion dollar platforms, complexity becomes inevitable.

Some developer user stories may remain simple (e.g. making a charge, sending an SMS) but a lot of the value will be created by systems integrators / service partners who understand the platform's products deeply.

Having good self-service products and documentation for the impatient developer is a necessary (but not sufficient) step to create this integrator ecosystem. Much of the work is on the business side to convince integrator managers that there is a viable market opportunity on your platform. This comes with all sorts of other considerations, like certification, co-selling agreements, reseller pricing, etc.

For all of their faults in product, Salesforce, Microsoft, and AWS have achieved their dominance through successfully building the integrator ecosystem. Hoping Stripe's service partner program takes off in the same way as the platform grows.

The AWS dashboard is like trying to learn Dwarf Fortress for the first time.

The number of companies (big ones, public ones!) where I can't for the life of me figure out what they sell from their website is truly astonishing. I understand that 1) if I can't understand it I'm not in their market, and 2) they probably want to be vague so they can sell their prospects whatever they ask for [1] but I still feel that they could gain a lot in the long term through brand awareness and people like becoming or meeting their target customer (but not having had any idea what that company did) by making a clear articulation of what the hell they do.

[1] Funny story, I was at a market in India and asked a guy selling toys and things if he had any fireworks. He asked me a lot of questions about what I wanted, but I couldn't see any evidence of him having such a wide variety of stuff. Then he sprinted off and asked his neighbor vendor to watch his stall. He came back 5 minutes later with a ton of fireworks to my liking, I bought what I wanted and then he went back to the fireworks guy with his leftovers. I assume that in a split second he negotiated a deal with that guy where he borrowed the fireworks, sold me what he could, and took the rest back. I don't know what he marked me up but I hope it was a lot.

Your "funny story" is very fitting here... there's a lot of value hidden in just knowing how to interface with AWS. Lot of companies are actually the middleman between you and AWS (e.g. Heroku which provides managed databases, Redis instances, etc. which are all hosted in AWS.) Meaning that you are not the client of AWS; Heroku is.

partly it is because in India you need licensing to sell fireworks , the demand is very very uneven, during Diwali season it is 10-30x compared to the rest of the year.

the fire hazard and pollution issues and till a while back child labour concerns as well is why it is regulated

The permanent year around sellers are few and serve as dealer than retailer . During the season they do both .

It was more sourcing a dealer on the fly and pretty common for fireworks .

Price is very fluctuating and there are huge markups due to the nature of demand .

Discerning buyers go to largest shop/wholesaler to get best price.

I knew a wholesaler/ early e-commerce player in this space (bigger players do not this due to regulatory challenges ). He will make his orders in January for delivery in august for season starting in sept/oct and typically paid 10%-15% list price at most.

Hi pc, huge fan of Stripe and its relentless focus on the "impatient developer", which so many enterprise companies miss the mark on. As a future co-founder interested in this space, do you have any talks or essays on how you built a developer centric company that has also a successful "enterprise" side to it?

Not pc, but they did an episode of "How I built this" that: https://www.npr.org/2020/03/06/812853591/stripe-patrick-and-...

I don't remember if they answered exactly what you are asking, but my memory of the episode is definitely positive!

they most certainly don't discuss enterprise go to market in an NPR podcast haha

Yes Patrick great work.

Having had to jump into a Stripe project recently, I think you may need to 'level' up on the experience as the number of features and products expands, it's getting more confusing.

'Everything is there' in the docs, it's just that there are so many now.

I don't remember the mappings of all the product names to 'what they do' (I should not have to) and I find myself getting confused whereas years ago I was not.

Everyone fails at this basically, but because you have a mission focused on this, it'll be interesting to see how you handle it differently.

There is not a single documentation search solution that seems to work: Apple apis, MSDN, Google Assistant, Stripe, Shopify, they generally can't answer basic questions and often don't point you in the right direction.

A question such as 'Can Stripe Checkout Client-Only Version offer Individual Discounts' - is a question a client had. It took too long to find the answer.

I suggest someone will some along perhaps with some AI and totally re-invent helping people find simple answers to simple questions, maybe you can lead the way.

Hiring Mike Clayville to Stripe is a huge deal, and a loss for AWS. This is one of those deals where both Stripe and Mike come out better for it.

>This is one of those deals where both Stripe and Mike come out better for it.

...is hiring not usually an event where employer and employee both benefit?

Often not, but it's considered rude to point out examples thereof.

Come on, stop making fun of Yahoo like that, it's been years now...

I seriously love the way Patrick Collison just randomly browses HN and takes the time to respond to articles like this.

I hope you give him more credit than "random"!

This is sort of neat because the same problem seems to be mirrored for Stripe customers; for example, if my business caters to the impatient developer and I use Stripe for payments, I find myself having to explore outside the bounds of conveniently orchestrated payments, to more esoteric per-customer high-touch interactions for large enterprise customers. I wonder if there isn't room for Stripe to grow to help businesses deal with these... sort of business deals!

Hah. This is in fact a problem area that our Billing (https://stripe.com/billing) team is actively thinking about. You should reach out to ark@stripe.com if you want to discuss -- he can connect you with the right folks.

There is a pretty simple feature that would alleviate a lot of pain with Stripe’s billing service when invoicing enterprises - allow for attaching a pdf of the invoice in the automated email.

Many enterprise A/P departments require it. The lack of this feature has prevented us from moving off our existing invoicing system to Stripe. I’m guessing we are not unique.

It's not quite the same, but you can get the PDF from the Invoice object and send your own emails.

It's not just enterprise - many small / mid size places are much much happier (myself included) with an invoice attached.

AWS gets this right (even though invoice is high level). You get something that can be quickly forwarded down / up and around and end up in the AP system / accounting system etc in good shape (ie, vs a 6 page PDF of an email chain). They just care that third party invoice is there in good shape with approval by whoever owns the expense line its hitting.

One thing that could help with acquiring both self serve and enterprise customers would be effective, B2B focused Facebook ads. I have a ton of experience with scaling these for other big B2B tech companies, email is in my profile if that sounds interesting.

Thanks for addressing this. Each of these points was an immediate worry as I was reading the post, nice to see you’ve already thought of them and addressed them.

Will you increase your rps limits? Surely they are a big blocker for larger organizations.

Hm, what issues are you running into? Our defaults (100 RPS) are pretty high but we can support higher if you reach out to us.

Reply on news.ycombinator.com