SaaS & the Art of Software Pricing. A decade of AWS & Twilio experience in one hilarious hour.

1 comment

A very sharp talk from the founder of Twilio on software pricing, including how to measure value and put a price on it. Jeff is also an expert on drawing owls and can teach you his patented method.

Jeff talks about the difference between cost based, value based and competitive pricing, as well as discussing the customer feedback loops, split testing, creating pricing options, audience segmentation, and using price levers.

Jeff co-founded Twilio with over 12 years of entrepeneurial experience, bringing product, engineering and business background to the company. Jeff was awarded a 2010 TechFellows Award for Disruptive Innovation, and was named number 18th on Business Insider’s 2010 Silicon Valley Top 100 List. Prior to founding Twilio, jeff served as an early Product Manager at Amazon Web Services, and led the development of Stubhub.com as its founding CTO.

Warning: Jeff does use some words here that might offend!

For more like this, visit the BLN’s talk library and find seven years of insights on building a great tech business.

Transcript

[PLEASE NOTE, DO NOT READ BEYOND THIS IF YOU ARE OFFENDED BY STRONG WORDS. WE HAVE TRIED TO CAPTURE THE ESSENCE OF ALL BUSINESS OF SOFTWARE TALKS AS FAITHFULLY AS WE CAN. THIS IS PART OF THE JEFF LAWSON PHENOMENON]

My name’s Jeff, I founded a company called Twilio. Twilio is as Mark said, AWS for telecom. And I’ll give you a little bit about my background. What I wanted to talk about today was SaaS. Um, a few different aspects of SaaS, but particularly how you price that SaaS and I thought about putting a nice ass up on the screen, ya know ‘cause sex sells and it would wake everyone up and then I figure “that’s tacky” then I thought I’d do it anyway. [Laughter]

Sorry. Alright, so, quick, let me give you my background so you know who I am. This is my fourth company that I’ve started, although prior to this I was also a product manager at Amazon Web Services.

So, started out in ’97, my first entrepreneurial endeavor was my lecture notes company during the dot.com days raised a bunch of money, blew it out to 200 campuses nationwide, then I was the first CTO of Stubhub.com the online ticketing exchange after doing that for a while, I went for a total change of pace and started a bricks and mortar company for extreme sporting goods – skateboarding, snowboarding, surfing and that kind of stuff. A total change of pace, then after that went over to Amazon, was product manager over at AWS, and after I left Amazon was looking for the next thing to start and realized that each of my three previous companies had wanted to incorporate communications into the web apps we were building and every time we were stymied because we said we don’t have the foggiest idea how to make a phone ring and so we started Twilio.

And Twilio is all about simplifying telecom and bringing it to the masses. And the idea is to open up this black box of telecom and make it accessible to the millions of developers out there who can now innovate in this industry that is not new but is certainly new to most developers because we’ve never played around with it. But! That’s who I am and that’s the prospective I’m coming from.

Let’s talk about the general world of SaaS and I start out excuse me, I’ve had a cold like two weeks ago and it’s still like hanging around. Starting out with who are you talking to? Who is the customer for SaaS? And that’s the most important thing before you start thinking about right as of course understanding who your customer is.

I like to sum up the notion of who a SaaS customer is with this picture that’s become something of an important meme at our company. Uh, it goes like this: how to draw an owl.

  • Step one, draw some circles.
  • Step two; draw the rest of the fucking owl.

[Laughter] [Applause]

So what does this have to do with SaaS? This goes to the notion of who we are as a company and who we feel our customers are as a company and we have a word for it and the word is “doers”. Right people who figure shit out and get it done. Right and to them and to me this is sort of a rallying cry for draw some circles and there’s no instruction book for the rest you just gotta go figure it out and draw the rest of the fucking owl. Uh and that’s what being a doer is all about and so when someone does something great and the figure out something at our company we say, “way to draw the fucking owl”. And we actually have little owls that we give out.

Stuffed animals, um [Laughter] yeah ten owls uh I can’t come to work anymore! Um and so who are this doers right? And I also visualize it as the kind of person who’s up late at night, trying to solve a problem right? When everyone else has gone to sleep, they’re the one who’s figuring out and get past a hurdle and to check in some code and to mark something as solved. And I think that’s very important to think about is what’s the mindset of the doer and the customers and why SaaS appeals to them, because largely it is about permitting innovation and permitting progress on the timeline that the doer is working on and when they’re ready and they’re ready to buy and they’re ready to figure out what you’re offering them, you’re there to do it. Um, and I’ll talk a little more about that.

But the doers are so important to us that we actually have ads on the internet just advertising the whole notion of doers. Um, and they link to a page on our website dedicated to doers. Um, and we list some of our favorite doers, uh, people who just picked up their tools and got stuff done and people we admire because they’ve done and built great things. You’ll notice Patrick McKenzie is up here. He’s gonna be talking to ya tomorrow. Uh but he’s one of our favorite doers as well, he’s a one man software machine. And, so this is really important, right? Picturing these people who are adopting SaaS because you make it easy for them to do great things for their company and um, typically it involves getting started easily, adopting easily and figuring out what you do quite easily. And this is every kind of SaaS; there is Enterprise SaaS, but I’m mostly talking about the kind that most people are familiar with which is you can get up and going and sign up on your own terms and get started. And to me, that is the magic of software as a service. Its low up front, easy to get started kind of model. Um, so that’s what we’re largely going to be talking about.

So, we’ve identified the doer as a customer but often time you think about while I’m building software that allows a business to do something or a B2B company not a doer, um, so part of what I’m here to say is that, you know, a doer is your customer. Um, and in particularly when you’ve gotta sign up in a form where you can get started and a credit card field that the enterprise or company is not your company. But really, think of it first as a person. People are your customers. And so thinking of who inside that bigger organization who’s gonna click that button is really important to a self-service, SaaS type model. Even if a large company ultimately is your customer, well you have to start somewhere and starting with people is how you do it.

If you think of a large enterprise as your customer, not gonna work because you’ve gotta target a single person and for us that’s represented by “Jazzy Chad”. Jazzy Chad is a customer of our, actually now an employee of ours but he started as a customer of our and it’s like you think about this one person, right and what are his hobbies? What are his skills? What hours does he work? What tools does he use? And appeal to him, and he will become the advocate for using you inside of say, his organization or with his peers. But you have to think about the individual person not just the enterprise or the big organization, that’s where you get the doers. Organizations and enterprises are not particularly known for their doing-ness. Enterprises are not known for getting shit done, but people are, right? So, thinking about the doer and targeting at them is really important, right? So now there are some questions, how do you talk to doers? How do you talk to a person who’s very self-motivated, who really wants to get stuff done and is working on their own timeline to get shit solved and moved on and essentially solved problems?

Right? So how do we talk to them? And a lot of them are very smart, very motivated and they’re also very skeptical of marketing BS. And so I think there’s a few unique ways that SaaS industry has figured out how to talk to doers and empower them. So let’s look at a few examples, right? There’s a few there are really great as far as SaaS companies, and um 37 Signals has a number of products, Basecamp being the flagship uh that many in here I’m gonna guess actually use for project management and things like that.  Um, we’ve also got CRM, uh, I threw Batchbook in here, CRM product Campaign Monitor, great for email delivery and marketing, Zendesk we’ve talked about before, great for customer service and I’ll throw another one in there as well. Just as examples. It’s interesting if you paid attention during those screenshots, you’ll see a few common threads that I’m gonna highlight now. And I call them the “mating calls of the northern modern doer”. And essentially, these are the ways that the companies in the SaaS business, in the SaaS realm communicate to potential customers that “we are a solution for you”. It is the mating call, it’s what the doers are looking for to adopt something.

So let’s look at an example, I’ll pick Zen Desk um, what are the mating calls that Zen Desk gives off other than a fat guy on the first page? First of all, you’ve got the tour. I’ll go into more detail in a minute. Second, you’ve got pricing and third you’ve got try it out, get started, sign up now, get going. Those are the three mating calls of SaaS companies. And I’ll look at Twilio and you’ll see the same ones. And they’re very common across all these sites so let’s talk about them in a little more detail. NP Um, so first of all you’ve got the notion of the tour. Um, this is really interesting, you may think it’s really obvious, I’m gonna show you what my product does. Now for most SaaS companies, especially UIs, there’s screen shots, people call it screen shot porn, uh, that really gets people interested and I’m like I totally get how that works and people excited about sexy looking screen shots. Uh, for Twilio we don’t have a UI, ‘cause we’re an API company so it’s kind of our documentation and code samples and things like that but either way there’s notion of a tour, right? And there are lots of different varieties of it there are all these different aspects. What they have in common is they actually tell you what the software does. That may seem retardedly obvious but in the world of enterprise software, they don’t tell what the software does. [Laughter]

They wanna keep it a mystery. There’s some high level things, but ya know if you really want to figure out if a software solves your problem you’re gonna call the sales guy and the sales guy’s just gonna tell ya that is solves your problem, whatever you say your problem is. Whereas SaaS companies tell you what their product does. Big deal. Okay next, well ya like what it does, how much does it cost? Again, SaaS companies tell you how much it costs and you the informed, intelligent doer can make a decision is what the software does, solves your problem at a price you’re willing to pay for it. It’s called making an informed buying decision. And doers are very good at doing that. Um, and, and, and SaaS companies put the pricing on the page, right? This may seem obvious, right? We see all the SaaS examples that we see in common life and you see a pricing model on the page yet, again, in the enterprise software world 10 years ago, it was very rare for B2B SaaS or B2B software period to have a price tag on it. Call a sales guy, he’ll tell you how much it costs, after he finds out exactly how much you can pay. Next, of course is the getting started. So a sign up button, right? Uh, uh you figured out this software can do what you want it to do, at a price that you’re willing to pay that makes sense for you. Great. How do I get started?

Now, the doers they’re at 3:00am, figuring out how he’s going to send a marketing blast tomorrow. He doesn’t want to contact your sales guy, he wants to sign up and get going with minimal friction. Now, there’s a lot said and studied about sign up forms, about how much information to take, um, about how much friction you should introduce or remove, based on what fields you ask for and how many of them there are. But the whole notion here, really, the common thread is that you actually let them sign up. Again, with enterprise software, there’s nothing to sign up for. All you can do is contact a sales rep who might call ya back in a few days if you said your budget was big enough, and your organization was big enough, and your buying timeline was soon enough. Otherwise, you’re stuck finding other solutions.

Right? So this is a key thing, right? It may sound obvious, but there’s example abound if you look out there where companies don’t do this, right? We boil this all down to one notion of “no shenanigans”. It was recently discovered if you trawled the Twilio website and piped it through GREP with the word ‘shenanigans’, and piped it into word count ‘-L’, you got an answer that was about 384 instances of the word shenanigans on our website. So, we’re big fans of advocating for this “no shenanigans” approach to business. Tell someone how much it costs, tell someone what your product does, and if that makes sense to them, let them buy it. This sounds like it’s obvious, but it’s not.

Here’s a vendor I had to deal with at my last company. My favorite part about this, “how to buy”. You know, my wife tells me I’m really good at knowing how to buy things. Especially if they’re from Apple. Um, you know we’re all really good at that we know how to take our credit cards out, we know how to buy things we’re very skilled at that as Americans and even other non-Americans. Um, but if a company has to re-teach you how to buy your product, my feeling is they are doing it wrong. NP So let’s look at how people come to adopt this stuff and I call this the doer funnel and also in our company known as, “Mommy, where do customers come from?” So it looks something like this: you’ve got an adoption funnel that starts at the top with your marketing activities. You’re trying to reach your target customer. Um, you’re gonna send them to your website. You’re an online business, that’s primarily how you market your stuff, I’m guessing. Especially ‘cause its SaaS so your whole idea is send them to the website. Now, you let them read information; what does your product do? You let them see the pricing. How much does it cost? If they like it, they click a sign up button.

Now, hopefully, chances are you’ve got some form of free trial, some way to get started with low risk, no credit card, or maybe a credit card your don’t charge it for 14 days, there’s a variety of ways to do this. But the whole idea is that you give them a low-risk way to get started um and then eventually they pay and that’s what the funnel looks like. And your goal is to get someone through the funnel in each step as quickly as possible through the process and this is called self service. This is what I’ve been talking about. The whole notion that there is no sales guy involved, you just get them through this funnel, and get them to the point when they’re putting in their credit card and paying you as quickly and easily as possible because you solved their problem. We call this the Consumerization of Enterprise Software. Right? So stuff that 10 years ago was a big enterprise sale with a sales guy and a 6 to 12 month sales cycle is now a few clicks and a credit card and you’re up and running in an afternoon. That’s the Consumerization of Enterprise Software. That is what a lot of SaaS is addressing. A lot of SaaS is taking things that were done in the past that were very expensive and very time consuming and making it now simple to get it up and running and get started.

Um, so in our space, we appeal to developers, and we sort of think of it as uh, we’ve sort of got all these software products out there and you say “don’t you love that SDK?” Does it come in Ruby? This is the buying process; literally this is how the developers in our world think. Right, and if I’m a Ruby developer and your API doesn’t have a Ruby variant then maybe you should move on to one that does. Um, this is a very consumer way of thinking. Right? It’s very much about making an individual decision at that moment. Do you have a product for me that’s going to solve my problem at a price I like? That’s how consumers think. Um, and so is that just sort of independent people who don’t have any money to pay you because they’re not part of the enterprise business? Well there are also doers inside of enterprises, inside of big companies. What they’re doing is they’re leading the charge for adoption of whatever you have inside their bigger company. Maybe it starts with them personally. And that’s how Salesforce started; individual sales people buying Salesforce. Soon, enough people were buying it individually and expensing it that the IT departments said, “you know what, we’re gonna buy a site-wide license”.

Um, maybe it’s a departmental solution for sending emails and a department signs up for a campaign monitor or whatever and it grows from there. But no matter how you do it, this applies, because your doer can be at a small company, can be an individual, can be an entrepreneur or can be playing that role of an early advocate inside of a big company. So the way that looks in this pipeline is that you get people to sign up and figure out what you do and figure out you can solve a problem for them, but if they’re at a big company, they but in their credit card and signed a click wrap agreement that they didn’t even read, that typically doesn’t fly at big companies so when you move to a bigger stage of adoption, you move from the self service you kind of bump it up to what we call at Twilio full service.

The notion is that if you’re at a big company or a Fortune 1000 or even if you’re not a big company, but you’re gonna spend a lot of money, you might need a little extra hand holding than a credit card form and a terms of service agreement that nobody read and uh that’s what we’ll do for you. That’s where our sales team comes in. So sort of this hybrid approach of self service and empowering people to make a decision but if they have a bigger use case or are from a bigger company, and they need something slightly different from you as a company, you can handle that as well. Alright, because you don’t want what are potential great customers to go away because, well you know, I can’t put this on a credit card so I guess I’d better find someone else. So you let people know that look, if you’re a bigger company we can handle you as well, just contact our sales team. Right? So it’s kind of the best of both worlds. You can contact a sales team if you need, typically if you notice this funnel it’s at the bottom of the funnel. It’s after you figure out that we solve your problem. Not the typical enterprise model where well first, let’s do the sales process and you cut us a check, and then we figure out if we actually solved your problem. Right, that’s how typical enterprise sales work. This inverts it, puts the proof of solution at the beginning and if it does solve the problem, then you’ve got a very fast sales cycle at the end. We’ve got some really productive sales people because the customer already decided they want it, it’s just a matter of the customer wanting a slightly different price terms, or getting payment terms or paying by invoice, or getting and SOA or 24 by 7 support. Something you may not give to self-service customers. NP That’s how we do it and it works very well as a means of capturing both the initial doer um in a small company market also all the way up through and including Fortune 500s who want to adopt the solution but need a slightly different way of doing business with you. Um, so that’s sort of about who these people are and why they’re adopting and how they’re adopting that I find really interesting as the stage for why SaaS exists and why SaaS is interesting and why SaaS has a foothold now in the world of business software, um, whereas 10 years ago it was a very different model and it creates huge opportunities for us as entrepreneurs to go and build great business that get started fairly inexpensively, where you can discover the customers market because you’ve got these self-service people who come in and are very measurable and very fast and their decision making, right?

With a click or the lack of a click you can learn a lot as opposed to a 12 month enterprise sales cycle where it took you 12 months to figure out your product wasn’t serving the market. So this is very cool and it’s very good news for us as entrepreneurs. But now the question I’m gonna leave the rest of the time for is “How Much”? How much should your charge, you should think about pricing, and what sort of look you might get if you give the wrong price.

As we’ve said before, pricing is hard. Pricing is not easy, but there’s a few different tools and a few different ways you can look at it to make sense of it and um, that, that, that um, help you out. The good news here, this is really good news, is a few simple formulas that, that you need to remember when you’re thinking about pricing. Um, so first the simple universal pricing formula. You create some value for your customer and any value for a business should be quantifiable. It may not be obvious, but it should be quantifiable in some way. If you charge exactly that amount, they have no reason to adopt you. You discount it a little bit, and that’s your price. So the whole magic here is figuring out how much value you’ve created and then of course how much you should discount it but this is a simple formula that gives you guidance on how your customer is thinking about your, um, offering. And what you’re offering them, and what they’re thinking about when they’re making the decision.

So let’s dive into this a bit more, right? So it’s all about value created, measuring that is key and that’s pretty hard and um you can talk to customers and there’s a lot of ways to do it. But let’s start out by saying that one thing that helps us measure the value created is by making an assumption and I’ve got an assumption that I’m operating with, I don’t know if you all have this assumption as well in your businesses but consumers or business. SaaS for consumers SaaS for business. Um, the bad news is consumers are not rational. If you need any more proof of that… [Laughter] It’s pretty bizarre how consumers make their decisions. However, the good news here is that business are pretty rational. At the end of the day, every decision they make about how to spend money is driven by two factors: either it’ll increase revenue or it will reduce costs. And these are both measurable value points that you can measure how much revenue am I increasing for you how much cost am I reducing for you? Now, is it easily measurable? Not at all. But if you can paint a likely picture either with a you know, total cost of ownership calculator or something like that, or just paint a picture in their head that one of these two things is likely to happen if they adopt what you’re doing, that’s how you sorta get your foot in the door and that’s how business will think about it, right? NP Now, of course if you’re doing something even in the world of business and you have no measurable value that you’re providing your customers that’s okay too you can just put ads on it, that’s what the advertising business is all about. [Laughter] Um, business customers have this framework for measuring value, right, and that’s pretty cool because that’s a basis for how we should go about making our pricing. So now you figure out, you’ve got some value so you can go ahead and be proud, and put price on your junk, uh, or your bits and pieces, uh, your SaaS, put a price on your SaaS because you’re proud, you’re giving value to someone and that’s the key thing, if you’re not doing that in the world you have problems. NP So, let’s think about a few models though and these things should ultimately start coming together in the middle, um, let’s think about cost based pricing. So if you guys aren’t familiar with cost based pricing, you think about your margin, margin is profit over price, profit is the price you charge, um, minus your cost of delivering it, so your cost is your price minus one over your margin. So you can essentially decide what the right margin is for your business, you can decide that I’m going to be a high margin business, and my product and my market better back that up, or you can decide you’re gonna be a low margin business and go for volume, and my product and my margin and marketing should also back that up. But you can essentially back into knowing your target margin if you have a business plan you probably have a target margin in there to come up with what you think your price in the market should be.

So let’s run through a quick example. If your target margin is 50 percent, that’s a pretty easy example, and your cost to deliver a service is 10 dollars, you do the math and your price is 20 dollars. Right? Pretty easy. Um, so cost base approach would say okay I know what my costs are and I’m gonna build in a fair margin and that’s how I’m gonna uh, determine my price for my customers. Right? So that’s one way of doing it. It’s a cost following model, that’s for example how Amazon Web Services goes about its things. You’d say, “Well why would you ever do it that way?” Well it’s sort of interesting; at Amazon Web Services they target market like storage, and computing, utterly enormous, enormous markets. And so they actually went with a low margin approach. Cost following. Take the cost to deliver that service add in a fair margin, but no a huge one and that’s how you price it. Why wouldn’t you decide on a huge margin? Well then someone else is gonna come in and undercut you. So add in a small but fair margin, but one that discourages competition. Um, and that way you’re going for sort of the core of the market you’re going to go for scale, and because you’re going for scale, you’re also going to try to operate at a smaller price structure than anyone else in your industry. So if you go to market with a low margin and low cost, you get more scale which lowers your cost, you then lower your price, and I know this morning we heard, “you shouldn’t lower your price” I disagree. You should lower your price if this is your pricing strategy and you increase your market even more, right? And it is a loop that keeps on growing your business.

Um, this is roughly how Twilio, this is how we look at our pricing structure as well. We are going for a very large telecom market that is utterly enormous, and we believe by lowering prices we encourage more and more use cases to get built, and get more and more people to adopt and build things that they wouldn’t have built at a higher price. Therefore, as we achieve greater and greater scale we pass those savings on to our customers. And indeed, more and more stuff gets built, right? And so that’s how we’ve approached the market. NP But it’s hard with SaaS [cough] what if your unit costs, your unit costs in that equation are basically zero? And this is actually the case with most if not all SaaS products because you’ve got two servers and a MySql server and your total costs on a monthly basis are about 150 dollars. Now for some amount of time, you can add a whole lotta customers to that configuration, right? Maybe add another redundant MySql server or another web server or something like that but essentially, you’ve got a fixed cost and you can keep adding customers to it for a very hard time so it’s hard to say, “Well, my actual cost to service that customer is 4 cents, that month if you do the math so therefore I’m gonna charge them 6 cents as per the cost following model. That’s probably not how you should think about it. So you’ve got value based pricing. Um this is more useful for business where the hard costs are very low, and the unit costs of servicing another customer are very low if not negligible. This is also a way for you to capture more value in your pricing than you would otherwise. So let’s think about this equation: start does my business generate value? No? Well you’ve got other problems don’t worry about pricing it, yes? How much value? Right? Back to that initial question, how much value am I generating? There are ways to think about this. Are they saving money by adopting what you’re doing? Great. Are they replacing a more expensive solution with yours? Great. Figure out how much, and you’ve go a price. Are they saving money because of efficiency gains? They didn’t use anything before, but they’re moving from say spreadsheets to CRM to manage their sales team. How much money are they gonna save by having efficiency built into their business? Or, are they generating more revenue because of what you’re offering them? The answer to all three of these is roughly the same: how much. How much are you saving them or how much are you making them? Apply some amount of discount and you should have a winning formula for the customers, right?

It’s pretty much that easy. Of course calculating the value is hard but talking to customers is what it’s all about. The last one, we’ve got cost based, we’ve got value based, and now you’ve got competitive pricing. So what are the competitors in the market charging and how do you fit into that? And I actually say for the most part, that’s a red herring. You should see the previous two slides to determine your pricing strategy. You can use this as some guidance; however, I don’t think competitive pricing is really a strategy. It’s a way of understanding where you fit in the market. So now what is interesting though is that now you can do a sanity check. You can look at all three of these models and if they converge in the middle, if your cost makes sense and your margin makes sense, and the value prop makes sense and from a competitive landscape point of view it makes sense, you’re probably doing something right. And that’s where these three different models kinda come together and can give you a sense of if you’re on the right track, right?

So then the question is, “is the price right?” We picked the same graphic for this, we picked the same graphic, do you have the right price? That’s a question that now; only interactions with customers will tell you. So you need some feedback lops, you need some ways of testing things with customers and seeing if it’s gonna work. So, I’ll give you and example: A few years ago, 2006 I think it was, uh, I had an idea for uh an API that would do the deliverability of email for you. So it was an SMPT gateway that you’d send through that would figure out how to get it to the inbox of Hotmail and Yahoo and Gmail and everybody else. NP I called it Mailspade, literally I bought the domain, it was good enough, I put together this jinky ass site, um and basically it explained the value prop in a pretty bad way and I put a price on it. And what I did was drove traffic to it, and this is one of the methods that Eric Ries talks about in the whole start up world is testing hypotheses and Google is a great way to do that it can drive traffic for very little money it’s like put up this site called Mailspade to see if there was any interest in this product. Then you do some kind of split testing and you drive traffic to it and so here’s an example of an email I would have been targeting or a search term I would have been targeting: transactional email API and you can see there’s actually companies now targeting that I think I was roughly the only one at the time. And you can do some split testing, right? You can have 10 dollars per month plus half a penny per email or 1,000 dollars per month and a tenth of a penny per email, or no minimum commitment a month and 10 cents per email. And just drive a bunch of traffic to and essentially what you’re looking for is who clicks that really jinky looking sign up button on top and how many people clicked it based on the different pricing models put out there. And you can learn a lot very quickly based on that simple little test.

And the tools are very robust out there now to be able to run those tests there’s a lot of AB testing SaaS companies out there, there’s people like Optimisely that let you do testing with drag and drop with ease so it’s very cheap to run these tests. Um there is the question so you run it on your brand if there is an existing company or do you create another fake brand to run it as? It is hard to run pricing tests on a live product, um but I don’t actually have any advice on that, so… [Laughter]

The key thing is, talk to potential customers um and learn the value prop during some private period when you’re still figuring it out and it’s not public and it’s not widely distributed and what’s interesting, we did this at Twilio we did this at the very first version of Twilio, we gave it to a bunch of people who said they would be likely customers at the idea stage, we went and talked to a bunch of people and found out that they would be interested in such a thing. So, we built an alpha version, didn’t scale, it was all running on one server, and it was half implemented by we gave them the docs and said go to town, give us feedback, right? And what’s interesting is it was free, we didn’t charge anything. This was our way of getting feedback from customers. We don’t charge during a private beta because we’re gonna take pricing out of the question, what we want is for customers to play around with it, hack around, see what kinds of things they can do so they can give us feedback and then we didn’t price it because then we wanted to turn around and get their feedback on the price.

Sp2: Do you eventually start charging those customers?

Jeff Lawson: Yes, we do. So when we turn it on, we start charging those customers, you’ve gotta give them a heads up, hey we’re going to be turning on billing next week, we really appreciate all your feedback, hopefully you got some value out of it, but we’re giving you a heads up we’re gonna start charging. Um, but the interesting thing is we want them to understand the value a little bit up front, then get their feedback on the pricing an um there was an interesting interaction I had with Mailspade so I ran this little experiment then I got busy doing other things, um, namely starting Twilio. I got an email one day from a customer, a potential customer, and he was asking me about the site and the functionality and all the stuff, right? And I actually got into an email exchange about what he would charge or what he would pay for it. Um, I don’t have the whole email thread but it was really interesting because he actually started out I got a call from another customer right around the same time who said “You know I couldn’t find any contact information on the site so I did a ‘Who Is’ and just called the phone number that was listed on the domain name.” Yeah, I would have figured that was enough incentive not to contact a business, if you have to do a “Who Is” to figure out who they are and talk to them, but some people out there are, are, are pretty aggressive.

So it was interesting right, I did get to speak with some customer early on in this experiment to ask that what are they looking for, what would they pay for it, how much value was I providing them, what were they using today etcetera and all those important questions. So in addition to having that sign up form, by the way the sign up form on Mailspade didn’t do anything, it said “coming soon”, but I did have a spot for you to say be alerted when this comes available and I took an email address, and the idea there was if I decided to move forward with this and decided to move forward to pricing, um, I had a list of people I could contact and I could get feedback on uh, the pricing. So you can drive the customers and collect email addresses and have a conversation with them.

Of course, now you’ve got SendGrid available who’s built a big business around it, good for them, and I have a highly-ranked SEO site if uh anyone wants to buy it and compete with them. Um, and so, uh what else? I’d say another goal so so you’ve come across what you think is the right price because cost based and value based and competitive based pricing and talking to customers has all given you an idea of what you think that pricing should be. You have an idea in your head, you have a proposed pricing plan. My advice to you is to start conservative. What does that mean? Overprice. Charge more. Because you ca always drop prices if needed. And I think dropping prices is a great way to delight your customers. You shouldn’t be tricky though, ha ha, um, you can, however, raise prices as well, but it’s very hard. Um, and as was said this morning, you need to grandfather people in, or else you’ll have a big outcry, I think that is very good advice but um but you can raise prices. And people who are continuing to get the old price will be happy, people who get the new price probably won’t even know, um and so you do have the ability to adjust you just need to be careful with your current customers and in some instances you can raise the prices on current customers, but it’s very tricky to do unless you’re skilled in the arts I would not advise it. Um, at least at this point.

But the key point is if you start conservative, you have room to lower prices and delight your customers. If you start really aggressive and start to increase prices, you’ve got a little more of a tricky situation to deal with on your hands, that should be obvious. So, you’ve got the price you went out on the market you were a little bit conservative, um and you marked it a little higher than you thought you could, just to be conservative. Do you now, have the right price? Maybe you’re still not sure. Maybe you don’t know. And it’s a temptation to add more pricing options. What if this part of the market doesn’t want to pay and that part does, and this feature costs more than that one, and coming up with more options just sorta makes sense, right? Figure out different ways and put them all up there and see what sticks on one pricing page. No! Wrong! Don’t do that! I believe in the world of SaaS, consumer products, groceries, you don’t want to do that. There’s something called the “Paradox of Choice”. The Paradox of Choice is if the consumer has too many options, you can read this book but I’ll distill it for you, too many options, instead of picking the optimal choice, consumers don’t buy. It is overwhelming. You overload their decision circuit and they move on. So literally, when someone is trying to figure out what to buy, and there’s too many options, instead of picking one, they buy mustard instead. Right, so you need to be careful in your pricing model. It’s okay to have a few choices, but it you overdo it, you can pay us monthly, you can pay us annually, you can pay us daily, you can pay us by product line you have this add on and that add on right? [cough] this will hurt your adoption rate because you will overload people’s decision circuit so keep it simple is really the lesson here. Um, and there’s science behind this, mostly involving barbeque sauce.

However, there are ways to do this successfully, and you see this a lot, audience segmentation. There are ways to take certain customers and segment them off from other customers and charge them differently, you just have to make sure you’re pretty disciplined about how you do it, right? So there’s a few properties upon which it’s very common to segment and I’ll walk through a few of them. The first is quantities; so the 37 signals guys do a lot of quantity based segmentation. A hundred projects with the premium, max is unlimited and plus is 35. So, depending what size company you are, how many projects you have [cough] is a pretty good proxy for that, so we’re going to try to target the pricing based on that. Um, there are other things too, right, that go along with it. Gigabytes of storage, number of users though they’re all unlimited, though so that’s a stupid one, but essentially, the idea is pick a few things that you think a bigger company with more with deeper pockets is going to spend and segment based on that. And that’s decent. Um, there’s another one here, right, you’ve got another one uh, uh, uh, which is features. Right, so features is another interesting one and I’ll pull up a pricing grid for this example. So why would you pick the uh, why would you pick the gold or platinum one? Well, because you get more features than you would otherwise. And that’s a reason why you would adopt. And in particular, what features? Well, let’s look at these. NP Dedicated IP, that’s good for deliverability, white labeling is good for agencies, uh, sub user panel again that’s good for agencies, they figured out their audience, agencies are big users. And in fact, they’ve got many customers underneath them and so they can pay more if we give them features that are specific to them that unlock new capabilities of the product. And another one is support. And this is one that we use actually, um so we charge more for more aggressive support. This is sales force’s model too; the major difference between this additions is 24 by 7 premier support. Um, and this is interesting, we do this too, we charge extra for 24 by 7 phone support, um, why? Because well, it costs us a lot more to offer that product. It makes sense that we would be able to charge more for it but also you look at the people who want or need that level of support, they tend to be big enterprises. Um and often times in a procurement department there’s a check box that says do we have 24 by 7 phone support, right? And they need to check that box in order to adopt you so you need provide it if you want that customer um, but it’s interesting because that is um another good revenue center for you, it’s a good segmentation method. And lastly, and I really like this one, hats off to Patrick, appointment reminder, uh, industry segmentation. So certain users of your products have more demands and might have more money. In this one, HIPPA compliance; check box at the end, look at that price differential! That’s a pretty good one, I like that one a lot.

So, there’s all sorts of ways in which you can try to segment your customers. Uh, again, knowing your customers will help you know how to do it, figuring the right vectors to turn on the customers you want, and also have an attractive price to those customers who don’t need those options, really critical. There’s another interesting thing going on in each of these pricing models [cough] there is a recommended, a highlighted, a one that draws your eye option that they want you to pick. And it is neither the cheapest nor is it the most expensive. Well, why wouldn’t they just put the most expensive one? As the everybody, click on $699? You must want HIPPA compliance? Well, there’s an interesting psychology of the human mind, but really, we should all have like, psychologists working on our pricing teams because that’s really what this is. You will want to offer less attractive options. So compared to that, compared to the max plan, premium plan is a screaming deal. Right? Thank God I’m not paying $149 a month! It’s only $99, right? So by positioning where you think most of your people are really going to pay, next to less attractive options, you can do that on both sides, you can say this one is way, way too much money! Thank God I’m not paying that! On the other side you can say that clearly, that one is for companies that aren’t as sophisticated as mine, um, you can get people to buy the plan that makes sense for them. Right? And it’s very interesting, right? Only suckers would buy those other plans! And I’m not one of those guys! So there’s an interesting psychology there and often times it’s called anchoring, right, price anchoring, so compared to the most expensive one, this is cheap. Right, it’s the same psychology that goes one when you’re gonna go buy a blender on late night television and they’re like $49.95, you’ll get this delicious blender! But tonight only, $12! Right, they anchor in your mind, “this blender is worth $50” and now I can get this $50 thing for twelve bucks! What an amazing deal? Who the fuck knows if it’s worth $50? They just told you that. It’s worth about 8 cents, right? So the whole notion of price anchoring could take someone who would not have spent very much because they thought I was worth 12 cents or someone who wouldn’t have bought it at all, but only ‘cause it’s a screaming deal does something jump out at them and says maybe I will get this because it’s my one time offer to get this going, right? And you also see words here. “Only” $99 a month, “Recommended” silver plan there’s a lot of things you can do to get that buying behavior working uh, in your favor.

So, to summarize here a bit, start off by knowing your doer and issuing mating calls make it so they are contemplating your pricing plan and make it clear that you have a service that appeals to them and will solve their problems for which they should pay you some price. Next, figure out, in some way, how much value are you generating for that customer? Somehow, you must generate value. You’re either gonna make them more money, or you’re gonna save them money, but you’d better be doing one of those two and if so, figure out how much. Test that assumption. If you assume I’m saving your $100 a month, test it, see if that’s true. Test if I’m saving you $100 a month, would you pay me $70 a month to save that a hundred? Um, betas, it’s a great way to test out assumptions with customers, make them free is my strong recommendation, and use that as an opportunity to learn about the value prop you’re giving to that customer, and get the pricing guidance you need from them after they’ve figured out the value prop.

Once you go public, though, do it with a price tag. Make money, be proud, put a price on your junk, go public with a price tag and adjust, if necessary, ideally you’re adjusting down, right? And lastly, these pricing levers, and price anchoring are good ways to segment your audience so to get them buying in a category that makes sense for you. Um, that’s what I’ve got about pricing your hot SaaS. I think we’ve got a few minutes left if there’s any questions. Do we? Alright. [Applause]

Sp3: Um, how much conflict do you see in the giving to much choice and the pricing options? Is there a rule of thumb for that?

Sp1: Conflict?

Sp3: Well, uh like how many different pricing options would you give where um, eventually it would become too much choice, right?

Sp1: Ah, you know I have no science behind this, so at like Twilio we don’t have a monthly set subscription pricing model, so it’s not an area where we’ve had to invest a lot of research, and anecdotally, I would say three or four is the right number. I um like three personally [cough] I think when you look at the 37 signals guys who have done a lot of blogging about this, and that pricing page, for them, and it should be for everybody, one of your key conversion pages, so you spend a lot of time optimizing it. They used to have, I believe, five across, um, and free was one of them. They went down to three, I think they had four maybe and now they’re at three and the free thing they had is still offered but it’s on the small banner underneath. Very much not the peer of those other options. So, I would look at what other people are doing but also for your business run the test. A B test it. Uh, but I think three or four is roughly what the industry would consider to be the right number. Anyone else?

Sp4: For products that are priced similar to Twilio or Amazon, um, we’re considering their price model and the one concern I have is that it’s not clear to the user as they arrive on the page how much the service is gonna cost them. Like so if we say we’re gonna charge you by the emails you send or we’re gonna charge you by DEC2 compute units it’s like that doesn’t amount to a figure in the user’s head necessarily. Have you seen that to be a problem or do you have any advice on it?

Sp1: [cough] Yeah, I think it depends on how clear the unit is to their business, um [cough] if it’s truly something they’ve got no, um they might currently use that they but they have like no real knowledge of it, having a calculator or some comparisons you’ve seen it on Amazon they compare their compute units to like, Zion processors. Pretty dated analogy at this point that they still haven’t updated. Um, but they sorta try to give you that comparison, um I think they might also have some calculator it’s like put in how many units you need, but that’s just doing math that’s not that hard. Um, the key thing is if you don’t know, and there’s a weird one um, I believe it’s their um, um, simple RDS, relational database service, where oh no, it’s the key value store one, anyone know what that’s called? Rd…Simple DB. Simple DB, they charge you compute units for how expensive your queries are but because you have no visibility in the underlying architecture, you have no fucking clue how good the performance it is! And so they had to go way out of their way to try and create some sort of calculator that would describe like if you had x-million rows and if you had this many indecises and did this and if it took 4 seconds, and I look at that and say that’s really way too complicated. Um, ‘cause clearing you’re mapping your pricing onto a model that your customers just aren’t thinking about. So what’s best, for us, our pricing units are minutes, messages, phone numbers. And most customers come to us and have some sense of what those things mean ‘cause those are real life things that us as consumers use when I spend a minute on the telephone I kind of understand what that is. Or I send a text message it’s fairly obvious, um, there are certain things you can do when it come to, especially if you have um, a, a, a competitive type of pricing you’re trying to compare apples and oranges, um to do a calculator to show what you’d spend if you use that model versus what you’d spend if you use our model. So in our world we have a lot of customers doing call tracking on top of Twilio. You provision a phone number, you put it on a web page, you can do analytics before you send the call off to say an advertiser. Um, and there’s been companies in the market doing that for the last five, 10 years. Charging a ton of money because it was completely opaque, no one knew how much it costs so you’d pay like $10 per phone number per month with us you’re paying $1 plus three cents a minute for the phone calls. So the big question that they buyer we found often didn’t know, how many calls does each of these phone numbers get? Well, we did a bunch of analysis, and we found that the average phone number in that use case when you put a phone number on a website, in like a directory or something, the average phone number gets like 7 phone calls a month for a grand total of like 20 minutes. So their $10 on us is like $1.10. Um, something like that it was a crazy comparison and yet the piece of information we kind of had to put together for them was, here’s the amount of calls that number is likely to get. Um, and once you know that number you plug it in and do like a calculator and you say, “Holy shit, this is a huge savings!” So, does that make sense? Done? Okay.

Sp5: Thanks Jeff! Well done, really, really good. [Applause]

One Response to “SaaS & the Art of Software Pricing. A decade of AWS & Twilio experience in one hilarious hour.”

  1. [...] SaaS & the Art of Software Pricing. A decade of AWS & Twilio experience in one hilarious hou… [...]

Leave a Reply