The collision of value-based pricing, customer dysfunction and ROI promises
Just because some companies act in ways which appear — and often are — wasteful, perverse and value-destructive, that doesn’t mean all their employees are blind to the inefficiencies. Ironically, at the heart of effective procurement, are decision makers’ profound understanding of the dysfunction inside their own organisation. It is the dysfunction that will prevent them from realising the full benefits of whatever they’re buying; but awareness of the dysfunction which gives them a truer understanding of the benefits.
“I would rather not have somebody put forth an ROI model that's insulting to my intelligence… I don't have it all figured out, but… an ROI model that's just fundamentally insulting to the work that my team does… turns me off hugely.” (Customer decision maker, discussing SaaS vendor ROI promises)
When was the last time a prospective customer expressed scepticism about your product’s ability to deliver on its claims? Not unease about the features, but doubts about the benefits. Disbelief that your proposition will deliver the promised time savings, efficiency gains, revenue enhancement, churn reduction or incremental customer acquisition.
Customer scepticism frequently stems from unhappy experiences buying products which under-delivered from vendors who over-promised. It is natural (and often sensible) for a prospective customer to assume the vendor is presenting the sunniest view of their proposition’s benefits.
However, as the vendor, it’s a mistake to assume that customer scepticism equates solely to cynicism about the vendor’s behaviour and incentives. Rather, a prospect’s pessimism is also grounded in the weary knowledge of their own organisation’s limitations and dysfunction.
This matters a lot to any SaaS or AI insights vendor using value-based pricing, and especially to those who combine value-based pricing with ROI claims in their sales or marketing materials.
My previous post started a series on the pricing, product and value challenges which B2B SaaS and AI insights companies experience in dealing with their customers’ execution gaps. This post continues with a focus on how customers perceive pricing and ROI promises.
ROI promises and value-based pricing
Value-based pricing is (or should be!) the default choice for SaaS & AI companies. Albeit, actually defining value-based pricing with appropriate customer segmentation, and validating customer willingness to pay does require some work!
A key challenge with actually selling the product is that customer perceptions of value will vary – often substantially. Even within the most thoughtfully implemented pricing strategy, with well-researched customer segmentation, there will always be a variance in willingness to pay.
So ROI claims are often used by B2B SaaS and AI companies to put some numbers on the benefits their proposition will deliver, and inherently justify value-based pricing. Moreover, prospective customers — especially those with formalised procurement processes — will do their own ROI calculations, and include these in their decision-making.
If the customer expects the proposition to plausibly deliver $N thousand in annual savings, and the annual fee is a fraction of those savings, then the customer has a positive ROI.
For example.
A. An AI insights vendor claims their product will reduce a telecoms company’s subscriber churn by at least 25,000 customers pa. With an average annual subscriber value of $360, the aggregate gross value of those 25,000 subscribers is $9m. The vendor proposes to charge $600k PA, effectively promising a 15x ROI.
B. A B2B SaaS vendor claims their product will automate administrative busywork and reduce customer support interactions, enabling their prospective customer to eliminate 2 administrative employee positions and 2 customer support reps. The fully loaded cost of these employees averages $60,000pa. The vendor proposes to charge $24,000 PA, effectively promising a 10x ROI
In both cases, savvy prospective customers will start to unpick the claimed benefits. But there are two parts to this unpicking:
The first is a sceptical eye on the “sunny” presentation of upside
The second is a weary acknowledgement of how their own organisation’s internal limitations will constrain the upside.
A sceptical eye on the sunny upside
If we look at the AI insights proposal, then the telecoms company will have some questions for the vendor! For example.
Did they get the $360pa ARPU figure from a random statista.com chart which wasn’t even accurate when it was published 3 years ago?
How much variation is that average figure hiding? If $360 is the mean, what is the median?
Where does this figure of 25,000 subscribers come from? What are the assumptions which went into it? And what does “at least” in “at least 25,000 subscribers” mean? Those two words are doing a lot of heavy lifting…
What proof can the vendor provide regarding the efficacy of their AI system for identifying subscribers at high risk of churn? What is the precision and the recall?
How does it compare to whatever we’re already using?
How will those customers be dissuaded from churning? What is the Retention Team supposed to do, and how much margin erosion is expected?
If there is net margin erosion, then how does that affect the ROI?
Does this require higher fixed costs in the form of more Retention Reps, and if so shouldn’t we be paying more attention to the net ROI?
So far, this is the standard procurement scepticism of any decision-maker who is familiar with vendors’ promises.
BUT… the savvier prospects will understand their org’s internal limitations and how these will drive down the ROI even more…
A weary acknowledgement of internal limits and perverse incentives
“Like one of those guys who buys a big new thing but doesn't really know how to get the most out of it!” (Toby Ziegler, The West Wing, Season 4, Episode 21)
Even if the AI anti-churn solution is remarkably perspicacious in identifying churn risks, the burden of actually saving those subscribers falls on the Retention Team.
This is where the decision maker starts thinking about how the Retention Team is incentivised, and how that might affect the volume and type of at-risk customers who actually get saved…
Retention teams are often incentivised via performance bonuses based on one or both of volume and revenue. If the bonuses are paid quarterly, and the churn-reduction initiative is driven via outbound activity, then Retention Team members are incentivised to focus on those subscribers with contracts expiring due before the next quarterly deadline. Those subscribers won’t necessarily be:
the highest value subscribers
the subscribers with the highest churn-risk
the subscribers whose churn-risk would be optimally targeted now rather than later. This point is important because a customer’s decision to churn is often taken long before contract expiry.
This is all compounded by staff turnover, which tends to be high in call centres. Staff are not incentivised to pursue churn risks whose contracts expire after the next quarterly deadline if they are less than confident they’ll still be working there.
The result is that staff are perversely incentivised to act in ways which will fail to maximise the ROI on the company’s investment in identifying those subscribers with the highest churn risk. The savvy decision-maker is aware of this and discounts the vendor’s ROI promises appropriately. In turn, this reduces the decision-maker’s willingness to pay because any solution is worth less than it would be if the Retention Team were more thoughtfully incentivised.
AI insights and conflicting goals
It gets worse! In other industries, the incentives can be even more perverse, especially when the incentives and objectives of different teams inside a company conflict with each other.
For example, some insurers are proactive in working with policyholders to reduce the probability or value of future claims. Health insurers incentivising policyholders to do more exercise via discounted Apple Watches & gym memberships is a relatively common approach. Notionally, this aligns the interests of policyholders (push back death, fewer diseases) with the insurer (less and lower future liabilities)
Moreover, ML-powered predictive healthcare SaaS solutions can identify individuals at high risk of making substantial future claims, and care teams can reach out to affected policyholders to discuss pre-emptive interventions.
But… what happens when the risk of customer churn overlaps with elevated risk of future insurance claims? Let’s say a defined cohort of health insurance customers have an expected average customer lifetime of 2 years. I.e. their insurer expects them to churn to a competitor in 2 years (e.g. because they change employers). A subset of people in this cohort are identified by the AI predictive healthcare solution as having an elevated risk of specific future health conditions or medical events which will result in expensive claims. The risk increases the further into the future you go, but can be managed with pre-emptive interventions (e.g. statins, incentives-to-exercise etc)
When insurers incur intervention costs today to reduce the risk or value of claims later, it’s because they expect — in aggregate — the future savings will materially outweigh the cost of the interventions. But if those savings are likely to be realised only after the expected churn date, then the insurer is hit with a painful quadruple whammy. They will have:
Incurred up-front costs for the predictive health AI SaaS solution
Incurred up-front costs for offering and implementing the intervention
Failed to realise any savings
Helped a competitor — to whom that customer will churn — by reducing the risk or expected cost of future claims.
This creates a situation in which the health insurer is perversely disincentivised to spend anything on reducing the risk of some of their policyholders getting ill. It also reduces the perceived value (to the insurer) of any predictive AI health solution. Yes, it might be possible to identify risks across a large number of insured people. But the insurer can only extract value from those predictions (in the form of reduced treatment costs) across a smaller subset of affected individuals.
This matters to the AI predictive healthcare vendor because, with a lower perceived value, there will be a lower willingness to pay on the part of the insurer, and so average contract value will be correspondingly lower.
It’s also easy to see how this can turn into a scenario in which different teams working inside the insurer end-up pursuing wildly contradictory objectives because their incentives are so conflicted. Does the insurer want to improve retention or get rid of customers with expected higher future claims? Does the insurer want to improve customer health more than it wants to avoid handing a benefit to a competitor? Does the Retention Team want to reduce customer churn or maximise their own performance bonuses?
This is all a great big bear trap for the vendor’s ROI calculation! The ROI calculation is premised on an insurer who acts to reduce their future liabilities. But, the insurer’s perception of those future liabilities, and whether they’re worth lowering, will be quite different depending on which team you speak to.
How ROI promises get discounted by prospective customers
This is a necessarily high-level list because there is so much variation in the detail across different domains and solutions. But it’s worth keeping in mind the ways a moderately savvy DMU member will discount a vendor’s ROI promises:
The vendor is presenting a best-case scenario, unmoored from reality. Prospect’s decision making unit (DMU) internally discounts the promised ROI.
The vendor has less/no visibility on the prospect’s internal metrics. The vendor’s assumptions on key variables are wrong. DMU internally discounts the promised ROI some more.
The vendor has lowballed internal implementation costs (or ignored them completely). Prospect’s DMU internally discounts the promised ROI again.
Vendor has ignored the likelihood that full deployment is only likely to take place after an initial proof-of-concept period or experimental deployments with a small % of a prospect’s customers. Prospect DMU internally pushes back the expected benefit timeline.
Vendor has ignored (or doesn’t understand) all the impediments, hurdles and friction identified in the insight-to-action loop. These will prevent the customer from gaining the full benefit of the proposition. Prospect DMU dials down the ROI.
These assumptions are important for three reasons:
They materially undermine the vendor’s ROI arguments.
They narrow the delta between the perceived real-world benefits of an expensive-but-advanced third party solution vs a cheap substitute solution (or no solution at all)
They illustrate self-awareness — on the part of the customer — regarding their own organisation’s inefficiencies and internal limitations.
They reduce the prospect’s willingness to pay.
What can vendors do about this?
I’ll be tackling this question in greater detail in a future post. But a key recommendation is to be careful in making ROI claims without consulting the prospective customer first. Not only is the claim likely to be wrong, but you run the risk of annoying your prospect, and denting your own credibility in the process.
It’s better to develop the ROI calculation in collaboration with an internal champion from the prospect’s org. Not only is this more likely to be accurate, but, as the vendor, you will accrue a much greater understanding of the prospect’s internal considerations and processes, and the context in which your solution will succeed or fail.
