Impact Pricing Podcast

Steven Forth is Ibbaka’s Co-Founder, CEO, and Partner. Ibbaka is a strategic pricing advisory firm. He was CEO of LeveragePoint Innovations Inc., a SaaS business designed to help companies create and capture value. Steven is what I consider one of the great pricing thinkers in our industry.

In this episode, Steven enlightens us about hybrid pricing models and explains why we should adopt it. He also shares why pricing people need to change the way they see AI along with the work they do in pricing.

Podcast: Play in new window | Download

Why you have to check out today’s podcast:

• Understand one of the unstated rules of pricing that talks about pricing power along with insights into future risk 
• Find out how artificial intelligence (AI) can help address concerns regarding the connection between pricing and predictability, and why pricing people should change how they see AI in line with pricing 
• Discover the reason on why most of us need to adopt hybrid pricing models and why the two pricing metrics in that should be independent of each other

“Most of us need to adopt hybrid pricing models, and the two pricing metrics should be independent of each other. If the two pricing metrics track each other closely, why bother having a hybrid pricing model? You need two metrics that are relatively independent of each other, and that will give you the flexibility you need to respond in a difficult economic environment.”

– Steven Forth 

Topics Covered:

00:42 – Steven’s insights on what a hybrid pricing model looks like 

02:30 – Issues on pricing model that applies credits/tokens to usage 

04:44 – Predictability; whoever has the best insight into future risk has the greatest pricing power 

09:37 – Pricing people vs. the recent trend: AI and how it sets prices 

12:02 – The Achilles heel of value-based pricing, and the hybrid pricing practice in companies 

18:09 – What real value-based pricing is about 

20:52 – An optimal pricing model if you want to combine two metrics 

24:31 – Confused buyers don’t buy and renew 

27:18 – Three reasons why calculators don’t expose their logic 

28:42 – Steven’s pricing advice 

29:42 – Connect with Steven 

Key Takeaways: 

“If you have enough data, you can actually get pretty good at predicting future usage.” – Steven Forth

“One of the unstated rules of pricing is that whoever has the best insight into future risk has the greatest pricing power. The better we get at prediction, the more accurate we can get at pricing, and the question then becomes, who has the data to make those predictions? And I think in many cases, it’s actually the vendor who has better access to data.” – Steven Forth

“This whole question of predictability and pricing is going to be a key question for pricing over the next three to five years, and the artificial intelligences are going to help us to answer that question much better than we’ve ever been able to answer it before.” – Steven Forth

“If you’re optimizing value for the customer, you’re optimizing the amount that they’re willing to pay you, so if you just do it to optimize revenue, you’ll end up shooting yourself in the foot because you’ll trigger that negative feedback loop.” – Steven Forth 

Connect with Steven Forth:

• Email: [email protected]  
• LinkedIn: https://www.linkedin.com/in/stevenforth/ 

Connect with Mark Stiving:   

• LinkedIn: https://www.linkedin.com/in/stiving/
• Email: [email protected]

Full Interview Transcript

(Note: This transcript was created with an AI transcription service. Please forgive any transcription or grammatical errors. We probably sounded better in real life.)

Steven Forth 

Most of us need to adopt hybrid pricing models, and the two pricing metrics should be independent of each other. If the two pricing metrics track each other closely, why bother having a hybrid pricing model? You need two metrics that are relatively independent of each other, and that will give you the flexibility you need to respond in a difficult economic environment. 

[Intro]

Mark Stiving 

Welcome to Impact Pricing, the podcast where we discuss pricing, value, and the hybrid relationship between them. I’m Mark Stiving and our guest today is Steven Forth. Well, you already know Steven; he’s been on here a lot. He’s a frequent guest, and he’s the founder of Ibbaka, and I’ve heard it pronounced both ways so I’m just going to pretend that I can say it anyway I want. Welcome, Steven. 

Steven Forth 

Mark, you’ll notice that in line with the theme for today around hybrid pricing, we have a hybrid name. You can pronounce it in two different ways. 

Mark Stiving 

As long as you pronounce it. 

Steven Forth 

As long as you remember it. As long as you can type it into a search engine. 

Mark Stiving 

There we go. 

So, we’re going to talk about hybrid pricing models today. Every once in a while, Steven and I will disagree or have a comment or say, “Wow, that was really interesting,” and I say, “Well, let’s talk about it on the podcast.” And so, we ran into some hybrid pricing models and we thought, “Hey, this makes sense. Let’s talk about it.” So, would you like to introduce the topic in terms of what do you think a hybrid pricing model looks like and why do we care? 

Steven Forth 

Sure. I think the simplest way to define a hybrid pricing model is it is a pricing model that contains more than one pricing metric, and does so explicitly. For example, you could have a pricing model that had a combination of per user pricing and a combination of usage pricing, and those two things are used together to create a hybrid pricing model. It’s sometimes used for pricing models that have complicated pricing metrics where you have multiple factors that go in to create one pricing metric. Oracle was the master of this; you may remember some of those complicated Oracle pricing spreadsheets where you had to put in the number of servers, the capacity of the server, and five or six other things and it ran a complicated formula that nobody understood and that spit out a price. You could probably get a PhD in mathematics studying the pricing of Oracle solutions, but that’s also another way of thinking about hybrid pricing. But as we’ve moved towards more transparency in pricing solutions, I think it is more commonly used to refer to having two or more explicit pricing metrics. 

Mark Stiving 

Let’s talk about the first one before we jump in, because I think the two pricing metrics is going to be more interesting and more fun. But if we think about the Oracle pricing method, and I was looking at a company called Databricks and they have data brick units, and it’s probably really similar, where essentially, you’re buying credits and they apply credits somehow to their usage. 

First off, that feels to me a lot like cost-plus pricing. And secondly, that feels a lot like nontransparent, and nobody actually knows what they’re buying or how much they’re buying, so I’m not a huge fan of that methodology. 

Steven Forth 

Yeah, it’s a super interesting question and very timely because if you think about it, that is how OpenAI is pricing access to the GTP language model, not ChatGTP but the actual underlying language models are priced per token. They have an input pricing metric that is based on tokens, and it is not immediately obvious to most humans how many tokens are in one of their prompts. I’ve studied it a bit and I can guesstimate usually fairly accurate, but it’s not obvious. 

There are advantages, though, to those pricing models. I don’t think they are necessarily cost-based; they can be value based. And the advantages are one, they can give you quite a bit of flexibility when you’re not sure how people are going to use things and what the pattern of use is going to be. And secondly, when you have a lot of complex underlying variables and you need to simplify and integrate the variables so that you don’t overwhelm the buyer. 

I think that there is a case to be made in certain circumstances for these sort of integrative hybrid models, but generally, I agree. I think that they have often been used to obfuscate the situation and we are in a world where buyers have less and less tolerance for that. 

Mark Stiving 

And I think the other thing is there’s no predictability. I’ve worked with clients who they say their customers get upset because they can’t predict how much they’re going to spend next month on our product. When we’re not selling something that we understand, then we can’t truly predict— “Hey, how much am I going to spend? How much do I need?” —which makes it challenging for our customers, which means it’s harder for us to sell. 

Steven Forth 

Yeah, either it’s harder for us or harder for them to buy. I think the way that that gets addressed is there’s a discount. Because there’s risk, they apply a risk discount. They may not talk about in those terms, but whenever there’s uncertainty in the amount of value that’s going to be generated or what the price is going to be, the price actually gets forced down because there’s a form of risk management taking place. 

However, Mark, I think that some of the predictive analytics technologies that have evolved over the last few years do give us the ability to predict usage and predict value in a much more compelling way. One of the ways that AI is changing pricing is things that were too unpredictable to be used as pricing metrics in the past are going to be usable in the near future as the prediction engines get stronger and stronger. 

Mark Stiving 

What’s a ‘for instance’ on that? 

Steven Forth 

If you have enough data, you can actually get pretty good at predicting future usage. Rather than just estimating future usage, you can actually put your data through a prediction engine, which is one of the things that AI’s do well—make predictions—and you can constantly improve the quality of predictions of future use. And the people that are doing this the most are people who need to do load balancing and Amazon and Google and Salesforce—anyone who’s running a large cloud infrastructure on behalf of others—are actually pretty good at predicting usage. 

Same thing is true in the electric power system. There’s been a lot of investment over the last decade and a half—two decades, really—on how to predict power loads, and we’re seeing this. We were talking earlier about the industrial internet of things and solutions in that space; one of the hot areas there is predictive maintenance. So increasingly, we can predict failure, we can predict use, we can reduce risk. And I think one of the unstated rules of pricing is that whoever has the best insight into future risk has the greatest pricing power. In the past, SaaS companies sort of forced their buyers to accept that risk and they thought they were being smart because they did that, but in fact, the buyers imposed a risk discount on them because they weren’t sure that they were going to get the value provided. 

So, the better we get at prediction, the more accurate we can get at pricing. And the question then becomes, who has the data to make those predictions? And I think in many cases, it’s actually the vendor who has better access to data. 

Mark Stiving 

Yeah. In my mind, I’m thinking of the company Splunk, and I know that Splunk is a very popular product, and one of their customers’ biggest complaints was they start using it and then people use it and use it and use it and use it and it grows super rapidly, and they end up spending a lot more money with Splunk than they ever thought they would, and they didn’t have a way to control it, to understand it. And I think that’s really the lack of predictability that I see from individual customers. 

Steven Forth 

Yeah. I think this whole question of predictability and pricing is going to be a key question for pricing over the next three to five years, and the artificial intelligences are going to help us to answer that question much better than we’ve ever been able to answer it before. 

There are simple things you can do as well. You can put a governor on the pricing, so that as the volume goes up, it automatically starts to make an S curve. There are ways to manage that. 

Mark Stiving 

I don’t really want to take this conversation this direction, but I just have to bring up one more topic, and that is black boxes. 

When we start thinking about AI, we’re kind of saying, “This is going to be a black box,” and yet, we’ve seen pricing systems go out over and over again and the biggest complaint is “I don’t want it to be a black box. I want to know why I’m doing this or why this is happening.” And so, I think these credits that people sell or buy are essentially black boxes that say, “I don’t really know what I’m buying, but I’m buying a credit.” 

Steven Forth 

I think that’s a fair comment on the state of the world today, but there’s a huge push towards what’s referred to as XAI for Explainable AI, and it’s actually getting written into legislation in certain jurisdictions, especially in Europe, and the US military has a big focus on XAI; it’s a powerful research theme. So, I think that we are going to get to places where we can do a better job explaining how the AI is working. And again, that’s going to be critical for people in pricing, because like it or not, most prices will be set by some form of AI within the next three years. And as pricing experts who are engaged with AI, we are going to have to push hard towards how we explain how the AI is setting prices. I’ve certainly heard people who use some of the old, heavy metal pricing systems complain that “The pricing software comes up with a price that they say is the estimated willingness to pay for this customer, but I can’t tell that to a customer. I can’t go to the customer. You should pay this price because my pricing AI said that you’re willing to pay it.” That’s not an argument that any salesperson worth their salt is going to make. 

So yeah, we have to change how we are thinking about AIs and how we’re able to explain what we’re doing with them. 

Mark Stiving 

Yeah, nice. Okay, let’s jump back to the topic we’re supposed to be talking about. 

Steven Forth 

Yes, hybrid pricing. 

Mark Stiving 

Let’s talk about hybrid pricing models. 

Steven Forth 

Yeah. 

Mark Stiving 

Before I ask you the hard question, can we agree in the following that if we charge for something, we get less of it? For example, if we charge based on the number of users, then companies try to keep the number of users from growing too rapidly. It’s kind of like a tax; if we tax something, we get less of it. 

Steven Forth 

I think you’ve identified what is often the Achilles heel of value-based pricing. The weakness of value-based pricing poorly designed is that it charges for some sort of input that is associated with value, and then because you’re paying for that input—as you said, use less of it—therefore, you end up getting less value. So, it’s possible in value-based pricing to have to design in, often unintentionally, but to design in a negative feedback loop. I think if we went out and looked at a half dozen or a dozen SaaS pricing models—SaaS is the area that I’m mostly focused on—I’m sure we would find a number of those. 

Let me give you an example. In the pricing and customer value management space, which is where my own software platform lives, there are four major vendors. Two of them do not actually publish their pricing; one of them prices per user. Now, this is an example of what you were talking about, right? If I’m paying per user, then I am asking my companies to reduce the number of users that I’m paying for. Therefore, the software will not get widely used within the organization and it will only be used by people for whom I can demonstrate value, because otherwise, why would I pay for it? If you’re not sure how it’s going to create value for a specific user or how a specific user is going to create value for the company using it, then the company will not pay for that user. 

Mark Stiving 

And I have to say what jumps out at me is if you are thinking that you’re using something like product-led growth, you’re absolutely hindering yourself here because you’re not letting other users try the product to see if they like it and have value to them. 

Steven Forth 

Yeah. And I would suggest that there are actually relatively few SaaS categories where the number of users is closely correlated with value. I think that in the SaaS industry, we defaulted to users because it was easy, and we needed some sort of metric so we used to do on license and we did it by per user or we did it per location, and location no longer makes any sense in the cloud world so we’re just going to do it per user, and I’m just going to write a simple formula that says they would have paid X dollars for the license and Y dollars for three years maintenance, so I’m going to take that and divide it by the number of months, either 36 months or 60 months, and voila, there is my SaaS price. 

Mark Stiving 

I often think that user-based pricing also came from salesforce, because they were so successful with it, everybody just said, “Let’s copy that.” 

Steven Forth 

Yeah. And let’s face it—an awful lot of SaaS pricing is copy my competitor who was copying someone else, and no one ever checked to see if any of this ever made sense, which I think is one of the reasons why usage-based pricing became such a hot theme over the last few years. And really, hybrid pricing is, in many cases, it’s a combination of user pricing and transaction pricing, or user-based pricing and usage-based pricing; doesn’t have to be. In many cases, it probably shouldn’t be, but if you actually go out and survey the different pricing models out there, that’s the most common case. You’ve got some sort of per user fee, and then you’ve got some sort of transaction or usage fee, and the hybrid is the combination of the two. 

But you know what? Even companies that say they are not doing hybrid pricing, most of them are. If you actually spend some time on their pricing pages, it’s hidden but it’s there, or it’s not represented as a pricing metric but it’s there. For example, HubSpot. HubSpot has hybrid pricing. It has a price per user and it has a price per contact. And the interesting thing if you look at Hubspot’s pricing is the price per user goes up from its cheapest package to its mid package and then it goes back down a little bit for its premium package. And there’s often an assumption that the price per user will go down across packages, but why do we assume that? In many cases, the value per user goes way up across packages, so why is the price per user going down? Because it’s what we’ve always done. 

Mark Stiving 

I think if you’re going to add more features at the next level of package, then it makes sense to increase price per user, because each user is getting more value or more capabilities—hopefully—with those features. 

Before I ask the really hard question, I want to go back to the usage thing then. If I charge you based on usage, so I’m going to charge you based on your HubSpot, and for every contact you put in, I’m going to charge you another $0.10. So now, you’re paying somebody to scrub your database of contacts. “These are people that we don’t talk to. I don’t want to pay my ten cents a month or whatever it is for these people,” so we’re incentivizing less usage. 

Steven Forth 

Yeah. Again, that’s an example of the Achilles heel of value-based pricing. When you design any kind of pricing model, whether it’s per user or for some sort of usage metric, you have to really think through what incentives you’re trying to set up. 

Now, the HubSpot one, I’m going to push back on that a bit, though, because first of all, it costs money to scrub a contact. It probably costs more money to remove dead contacts than it does to pay for them. And we’re a HubSpot user, so I should know this, but I’m not sure that I do, but HubSpot should be providing a service where it offers to scrub dead contacts for you. And I think that the companies that take that sort of step and say, “We are going to have a value-based price metric”—in the case of HubSpot, contacts—“and we are going to make sure that you get value from each of those contacts. So if you’re not getting value for that contact, you’re not using it, it’s a dead contact, we’re not going to charge you.” 

Mark Stiving 

Yeah, this is a lot like Slack. Instead of charging per user, they charge per active user. 

Steven Forth 

Yeah. And do you remember? I’m not sure if you’re a big Slack user, but I remember the first time I got a message from Slack saying that our activity had gone down and they were giving us a credit for the month. And man, they bought so much loyalty from me by that one email. I never thought about it again. I thought, wow, Slack, great company. I believe in them. 

Mark Stiving 

Nice. 

Steven Forth 

How do I give them more money? Because I’m sure that they’ll give me more value for each dollar I give them. 

Mark Stiving 

So essentially, what they’re saying is “I don’t want to just take your money. I want to take your money when I’m giving you value.” 

Steven Forth 

Yeah, and that’s what real value-based pricing is about. 

Mark Stiving 

Yeah. Okay, so now, not really a hard question, but when you combine the two metrics, so let’s say I want to do HubSpot pricing, and I want to charge per user and I want to charge per contact, what do you recommend? What does it look like? Do I actually have a separate price per user and a separate price per contact? What do you see as an optimal pricing model here? 

Steven Forth 

That’s one of those questions that is easy to ask and hard to answer. I believe that the answer to that is found in the underlying mathematics, and that you need to actually model that out under different numbers of users and different numbers of contact assumptions and build a model for that. And as an aside, I would suggest it’s time that we move from data informed pricing to model driven pricing. 

And the way to answer your question about HubSpot—and I’m sure that HubSpot has done this—is that you build a model and you explore the space that that model creates. There’s two approaches; one is you can either look for sort of large areas of similarity and use that to build packages or you can have very flexible pricing which is fine tuned to each individual customer. In Hubspot’s case, I think they’ve looked for the sort of large areas where customers are similar and use that to build packages. I suspect that we may see in the future a much more dynamic approach in which the amount that is due to, say, users versus contacts, to take the HubSpot example, will change dynamically in a way that is designed to optimize value. Because if you’re optimizing value for the customer, you’re optimizing the amount that they’re willing to pay you, so if you just do it to optimize revenue, I think you’ll end up shooting yourself in the foot because you’ll trigger that negative feedback loop that you identified. 

Mark Stiving 

In your answer, you implied the answer that I usually give when I’m asked a question like the one I gave you. And so, I’m going to add two things to what you said. 

First thing is I like having tiers or packages of usage. So even if you’re going to charge per user on HubSpot to create a good, better, best package with up to 10,000 contacts, up to 100,000 contacts, up to whatever the numbers are, what I like about that is it doesn’t disincentivize me growing my contacts. It’s like, oh yeah, put more contacts in, because I’ve got up to 100,000; we’re nowhere near that. And so now, we’ve got lots of usage and lots of incentives going on. 

And the second comment that I wanted to add to what you said is, I think as long as we’re selling without a direct salesperson, so if I’m expecting you to interact with my website and be able to figure out what you want, then I’ve got to keep these packages relatively simple and understandable so that people can make those decisions because confused buyers don’t buy. 

Steven Forth 

Yeah. I’m going to push back on that a little bit, though, Mark, because I think that the packaging approach…Ibbaka spends a lot of time designing packaging for our customers so I like it when people want to package; means more money for me. But packaging only works if you can find these groups that are similar. You build the model, you run the range of different scenarios through the model, and you look and see if there are similar groups. Usually there are, but it’s not a given. And as we move into more of these AI-based applications, it’s not clear to me that it will remain easy to find these packages of similar users. 

And then the second, on the ease of buying, I think that one can invite a person to input two or three pieces of data, and the software calculates a price for them, and it can even show the equation it uses to calculate that price to be transparent. 

That’s actually, in some ways, an easier user experience than it is for someone to go, “Okay, do I need the standard package or the premium package? When am I going to need to upgrade? What if I fall in between? What is the incremental cost of buying additional contacts if I’m in the middle package versus the premium package?” We can actually make package choice quite complicated by accident. And that’s why I think most packages only identify one pricing metric. In fact, they’re based on several pricing metrics, but they only identify one, and they vary the others, because it’s to make it easier to buy. 

But then we have to remember we’re in the subscription business, so we not only have to make it easy to buy, we have to make it easy to renew. And sometimes, the designs we create to make it easy for people to buy actually make it harder to renew. And so, I agree with your general principle, which is let’s make it easy to buy, let’s make the price transparent and easy to understand, and let’s make it connect to value. In some cases, I think packaging is the right solution to that problem, but not in all cases. 

Mark Stiving 

Okay. I’m sure that you can come up with an example, but I can’t think of a quote calculator that I’ve seen that makes me feel comfortable. 

Steven Forth 

I think that’s because they don’t expose their logic. They really are examples of black boxes, right? Either both the ROI calculators and the quote calculators. And the reason is that they don’t expose the logic. 

Now, some people are not going to be interested, but you should always have the option of asking how was this calculated? And I think there’s two reasons why people don’t share that; well, three reasons why people don’t share that logic. One is that it never occurs to them to do so, second is that the logic sucks and they don’t want to show it to people, and third is they’re using some sort of complicated AI and they can’t explain it, and that’s where your earlier comment about black boxes comes into play. I think all across AI, there’s going to be a huge demand for transparency, and in general, we don’t know how to deliver that yet. 

Mark Stiving 

Yeah, cool. 

Steven, we are running out of time. And even though I’ve already asked you this question three or four times in the past, I’m going to ask it to you one more time. What’s the one piece of pricing advice that you would give our listeners that you think could have a big impact on their business? 

Steven Forth 

Most of us need to adopt hybrid pricing models, and the two pricing metrics should be independent of each other. If the two pricing metrics track each other closely, why bother having a hybrid pricing model? You need two metrics that are relatively independent of each other, and that will give you the flexibility you need to respond in a difficult economic environment. 

Mark Stiving 

That was a fabulous answer. Would I expect anything less? 

Steven Forth 

I had it canned ready, too. I was just waiting for the chance to say that. 

Mark Stiving 

Okay, good. Thank you for your time today. If anybody wants to contact you, how can they do that? 

Steven Forth 

The best way to find me is [email protected], and I am very easy to find and very active on LinkedIn, Steven (with a v) Forth on LinkedIn. 

Mark Stiving 

Alright. And to our listeners, thank you for your time. If you enjoyed this, would you please leave us a rating and a review? You can go to www.ratethispodcast.com/impactpricing. And if you have any questions or comments about this podcast or pricing in general, feel free to email me, [email protected]. Now, go make an impact.