Value-based pricing is based on the theory that the more differentiated value you provide the higher you can set your price. There are some caveats of course. Pricing power is the degree to which you actually capture that differentiated value into price. The three critical concepts here are differentiated value (how much value is created for a customer), pricing power (how much of the value you could capture should you choose to do so) and pricing strategy (given your differentiated value and pricing power, how much of the differentiated value should you try to capture). The relationship between these three concepts is shown below.
It this post, we are concerned with one of the most important determinants of pricing power, risk, or more accurately, perceived risk.
There are several dimensions of risk to be concerned with.
- Supplier risk
- Customer risk
- Value risk
Supplier risk is simple to understand. Buyers do not want to invest in solutions where the supplier may disappear or the solution is no longer to be supported. This is a real risk, and not just with start-up companies. I have built several solutions on Google apps that are no longer supported (anyone remember Google Notebook?) and I was a big fan of Groove before it was acquired by Microsoft. In an era of cloud services, this is an even more serious risk as a cloud vendor can disappear and take your data with it but many of the most important innovations come from smaller companies and anyone who wants to be on the top of their game needs to use them. Buyers compensate for supplier risk in many ways. They may insist on code escrow agreements (where they have access to the code if the service is shut down), data warehousing (where they are able to download and store their own data) and service continuation and service migration agreements (to cover mergers and acquisitions). They also impose price discounts on early-stage vendors.
It is not just suppliers that change. Customers change business models too, and generally want to avoid lock-in. They are also afraid of paying more than the next person (especially if the procurement organization is involved in the deal). Smart customers (and we want our customers to be smart as that way they are more likely to be successful) are aware of the risks inherent in their own business and factor them in to buying decisions. There is not a lot that suppliers can do about this, other than be aware of it, and be flexible in helping customers manage their own risks.
The most important risk impacting pricing power is not supplier risk or customer risk but value risk. Frankly, buyers are sceptical of seller value claims. If there is only one value claim they look at it very carefully and test each assumption, discounting where there is uncertainty. When there are too many value claims, they are also sceptical, thinking that the seller is just giving them a laundry list and has not worked to understand their business. Generally keeping to two-to-four well documented and supported value claims is the best approach. Support the value claims with white papers, value calculators and data, lots and lots of well presented data.
Even with good reasoning and supporting data doubts will remain, and those doubts will get translated into a risk discount. How much of a discount? A naive calculation, one that many buyers actually make, can look like this.
Suggested Price $20
Perceived Risk of Getting Full Value 50%
Willingness to Pay $10 (50% x $20)
Simple to the point of silliness? Yes. But I have seen many customers perform this calculation. I have even seen salespeople offer it up. Ouch.
There are more sophisticated approaches, but the results are not necessarily any prettier.
A Bayesian Approach
Some years ago, I was working on a large deal to sell value-based pricing software to a major industrial company. One of the stakeholders was a crusty old CIO who said something that stuck with me.
"It really pisses me off when you vendors come in here and say your software is going to deliver some result when it is really the hard work of my people using your software that gets the result."
The truth hurts sometimes, but he is right. In most cases, it is not the software but how the software is used that creates the value. And in almost all cases there is some other way to get to that value. What does this mean? Let's try a simplified Bayesian analysis to see if we can get some insight.
Bayesian statistics are taking over. The highly popular FiveThirtyEight blog by Nate Silver and his team has popularized their application in sports and politics. And many of the more advanced AI systems (including the one used at our sister company TeamFit) have Bayesian networks working somewhere in the background.
The basic Bayesian equation is deceptively simple. P(A|B) is the probability of A given B. P (B|A) is the probability of B given A. P(A) is the underlying probability of A and P(B) is the probability underlying probability of B. Putting all this together the equation looks like this.
P(A|B) = P(B|A) x P(A) / P(B)
What does this mean to the risk discount in pricing?
Let's imagine that we are talking to the CIO. He has noticed an increase in sales in the market and he is wondering if this increase is due to his competitors adopting your software or to other factors. (We are going to keep the analysis here as simple as possible, I can already hear people telling me this blog post is too long.) Let's say there is an 80% chance that if this software is adopted, it will increase sales (you have evidence for this from other similar companies). How can anyone resist this! Well, they can and do, because they know that some companies are getting the sales increase without using our software. There is a 20% chance that sales will go up among companies that do not adopt the software.
Our question, given that a company is using our solution, and that sales have indeed increased, what is the probability that the sales increase was a result of using our solution, or what is the P(A|B). This is the information we want to get to the CIO.
To answer this, we need to know the value for P(B|A). This gets a bit tricky, so we will use a decision tree to help us. There are four possibilities to consider.
1. Our software was used and sales went up (80%)
2. Our software was used and sales did not go up (20%)
3. Our software was not used and sales went up (20%)
4. Our software was not used and sales did not go up (80%)
Let's put these into a simple decision tree.
We want to plug our numbers into the Bayesian equation to see what is the probability that an observed sales increase is due to the adoption of the software.
P(A) = .2
P(B) = .32
P(B|A) = .8
Plugging these into the Bayesian equation, we get the P(A|B) = .5
P(A) is the probability of the sales software being used, which is given as .2.
P(B) is the probability sales are up, which we can see from the Decision Tree is .32.
P(B|A) is Path 1 (Use Software and Sales Increase) is .8.
Plugging these numbers in we get the following.
P(A|B) is .5.
1) 50% of the companies who saw a sales increase used the software vs. only 20% overall and only 16% of those who didn't use the software.
2) The comparison of P(B|A) 80% to P(B|~A) 16% (those who saw a sales increase while using the software vs those who saw a sales increase without using the software) is even more compelling. We increase the likelihood of increasing sales by 5 times(!) by choosing to use the software.
If you can explain this to the CIO ,you will have a path to a shared understanding and it may be easier to find a way past this. Let's not forget, 80% of companies using the software see the benefit, while only 16% of companies that do not use the software are seeing the benefit.
Risk Adoption and Pricing Strategy
Being able to understand and explain the risk of getting the intended value is foundational to value-based pricing. There is an explicit or implicit risk discount built into the price that buyers are willing to pay, even when they believe the value propositions.
There are two ways to address this.
One can accept that this risk exists and that the customer bears the risk. Under this strategy, the goal is to get as explicit as possible about the reasons the customer will not get the intended value and the other ways they could get the value and then help the customer reduce the risk. This is often done by the people in customer success, who are frequently accountable for renewals, the assumption being that customers are more likely to renew if they are getting value.
The second approach is to shift the risk from the customer to the vendor. This is known as performance-based pricing. As the customer is no longer carrying the risk it does not need a risk discount and much higher prices can be achieved. This works when (i) the vendor is in a position to reduce the risk through its own actions or (ii) when predictive analytics can be used to clarify the risk and to manage it. Predicting risk will depend on who has access to the data and tools to measure predictive value.
So when you are developing your pricing ask the following questions.
- What is the value risk?
- How can the value risk be reduced?
- Who is in the best position to predict this risk?
- Who is willing to accept this risk?
The company that can best reduce the value risk or predict the risk should be the company that benefits. If this is the buyer, there is a risk discount. If it is the seller, there can be a performance premium.
At Ibbaka, we are happy to help you work through this for your own markets and solutions and help you to design the pricing model that will maximize your returns.