top of page

Expected values, decision trees, and why your customers are not purchasing your (or any) product


When reading Josh Bersin’s recent whitepaper, “People Analytics, Evolved: A Systemic Approach”, I was taken aback by one of the three options laid out for companies regarding the possible course of actions to tackle the massive data management problem. That option was “You can do “business as usual.” They added, “This is what most companies today do.” After a super clear and compelling picture-painting of the problem, how could an organisation choose to continue as-is?


Fast forward a few weeks to SaaSiest in Amsterdam, I found myself listening to April Dunford’s stellar presentation on positioning. She said, “40% - 60% of B2B sales purchase processes end in no decision. The easiest and lowest risk decision is to do nothing.” She continued, “buying a product is hard; and sales representatives putting their customers through the windtunnel of features are not helping.” That’s when it clicked: this is simply another math theorem at play. It’s buyers’ overly complex decision trees keeping them stuck in that indecision limbo.


Which brings us to the fourth edition of Math meet B2B SaaS, where I’ll provide a systematic approach to simplifying your prospects’ decision trees so that the answer to them seems as crystal clear as it does to you.


‘Math meet B2B SaaS’ is a mini-series where I discuss how math theorems can help explain and address some strategic as well as operational topics in the world of B2B SaaS.


The first edition of "Math meet B2B SaaS" covered how Poisson processes can help you think about your GTM strategy from a new perspective.


The second edition covered how you can leverage the optimal stopping theory to land on the best Ideal Customer Profile (ICP).


The third edition covered how you can leverage Bayesian statistics to ensure your Go-to-Market Strategy stays adaptive and data-driven.


Let’s start with the basics: What’s a decision tree?


A decision tree is a map of the possible outcomes of a series of related choices. It helps people make decisions by breaking down complex choices into a simple, organised structure.

Decision tree

Root node: This is the starting point of the tree and represent the initial decision or question.

Decision nodes: The nodes represent choices or decisions you can make based on certain criteria or conditions. They have branches leading to other nodes.

Leaf nodes: These are the endpoints of the branches and represent the final outcomes or solutions.

Tree depth: The depth of the tree is the number of levels or layers it has, with the root node as level 0. So in this simple visualisation, the depth of the decision tree is 2.


A decision tree is like a flowchart that starts with a question or a decision and then systematically explores different options and their consequences.


  • At each branch, you consider the possible outcomes or consequences associated with that choice. This may involve weighing the benefits and drawbacks of each option, as well as the associated costs.

  • You assign probabilities to different branches to represent the chances of each outcome happening.

  • As you move through the decision tree, you calculate the expected value of each option. This involves multiplying the value (benefits or gains) of each outcome by its probability and summing these values. This helps you compare and prioritise options.

  • Ultimately, you reach the end of the decision tree, where you have explored all the possibilities and evaluated the expected values.

  • And finally you can make an informed decision based on which option has the most favourable expected outcome. Ta da!


How’s this relevant to B2B sales?


Buying a B2B SaaS product is a complex decision, which can be simplified by building a decision tree. And as a sales lead, whether you build this decision tree for your prospect and guide them through it will determine whether you lose to or win against your biggest competition. And that is indecision.


Here is why:


Most sales leads initiate a sales conversation with a product demo; which means most start the conversation at one of the leaf nodes – the endpoints of a decision tree that you should be arriving at after having gone through one or more decision nodes.


In the case of B2B software purchase decisions, your initial decision nodes could be decisions such as ‘do nothing’, ‘build a system internally, ‘upgrade to a feature/module that a more end-to-end solution offers’, or ‘buy a next gen point solution’; whereas the leaf nodes can be specific vendors to ‘buy’ decision nodes, and various tools and talent pools to ‘build a system internally’ decision nodes.


By jumping straight into a product demo, you haven’t even discussed which decision nodes lead to these final outcomes, or worse, whether the prospect agrees with you on a certain decision node having a better expected value than others, before throwing them into the depths of a leaf node.


Most sales leads do this with the best intentions and motivations: they are passionate about the product they’re selling, they are super familiar with the problem statement and are already convinced about the case for purchasing their product, so falling victim to the curse of knowledge, they want to ‘wow’ their prospect as soon as possible. However, showing feature after feature to a prospect very early on in the conversation generates more confusion, which only adds complexity to their already complicated decision trees, rather than simplifying the decision making process for them.


Put simply, it can be likened to opening more and more tabs in your browser.

My brain as too many tabs open...4 of them are frozen and I have no idea where the music is coming from...

Here is how:

Whether people realise it or not, ‘do nothing’, ‘build’ and ‘buy’ decisions are very common initial decision nodes for prospects in a long list of B2B SaaS categories. These include accounting, salon management, workflow management, collaboration and productivity tools, revenue intelligence, customer onboarding, spend management, contract lifecycle management, process automation, cybersecurity, compliance, salary benchmarking, to name a few.


So let’s pick an example, say you are a company selling a customer onboarding solution, and look at the decision trees of two prospects:

  • Prospect 1: one who has been thrown deep into a product demo in the initial interactions, and

  • Prospect 2: one who has been guided through the decision tree to a leaf node (i.e. your product).

For the purposes of this exercise, let’s assume that both prospects share similar characteristics, and fit your Ideal Customer Profile (ICP) perfectly:

  • They have 5-person customer success teams.

  • They have 100 live customers, and lose 5% of the customer base on an annual basis due to poor onboarding.

  • Their average annual revenue per account (AARPA) is £20k.

And let’s assume that your product costs £100 per customer success team member per month, and can decrease customer churn due to poor onboarding from 5% to 1%.


Here’s what the decision tree for Prospect 1 looks like:


Prospect 1

You’re taking Prospect 1 directly to the vendor selection stage, where you are talking them through each feature your product offers. In their mind, they are battling with new thoughts and questions related to what each feature would mean for them from an operational and financial perspective.


Let’s now look at the journey that Prospect 2 is going through as they are guided through the decision tree:


You start the conversation by explaining the options available to organisations to tackle this specific problem. So at this moment, you’re not ‘selling’, but sharing knowledge on the market.


You start with saying something like, “Most companies today don’t use a customer onboarding solution, and try to manage the process with emails, spreadsheets, and video calls. This leads to roughly 5% logo churn due to poor onboarding, so the expected cost of not addressing this issue is c. £95k in ARR this year. This will only increase with an increasing customer base.”

prospect 1 continued

Notes: Top branch: 100 customers * 5% churn * £20,000 AARPA = £100,000 in lost ARR

Bottom branch: By chance, customers don't churn despite poor onboarding


Moving on to the second option:


“Some companies choose to build a system internally. Unfortunately, we know from research and having onboarded many customers who initially tried to build a system internally that 80% of those projects fail after incurring c.£50k in development costs. Not to mention the opportunity cost of not having that team focused on something else for the duration of build. So not many initially realise that the expected cost of building a system internally is c.£102k (excl. opportunity costs), which is a worse expected outcome than that of doing nothing.”

Prospect 2

Notes:

Top branch: Assuming £50k in development cost + EV(do nothing) if the project fails

Bottom branch: -£50k development cost depreciated over 5 years + (4 * £20,000) in ARR from reduced churn each year


Finally, the third option:


“Some companies use a next generation point solution to onboard their customers. These companies are those who care deeply about customer service, are keen to get their customers to that ‘aha’ moment early on, mostly during onboarding. These companies also tend to have highly efficient product and tech teams churning through a long list of feature releases, so the opportunity cost of deploying them on an internal build project is too high. Do these resonate with you? Great! So our experience suggests that with a next gen point solution, churn due to poor onboarding decreases to as low as c.1%. This would suggest an expected improvement of c.£39k in the bottom line, considering no upfront development costs that you would face with an internal build, and including the average cost of a next gen point solution.”

continued

Notes:

Top branch: -£6k product cost (5 CS members * £100 per member * 12 months) + EV(do nothing)

Bottom branch: -£6k product cost + (4 * £20,000) in ARR from reduced churn


The full picture:

final decision tree

Now your prospect’s decision has become a lot more simplified: you’ve essentially closed some tabs, without simultaneously opening new ones. You’re aligned with your prospect that buying a point solution would be the best option for them. You have successfully taken them out of the indecision limbo!


So what’s next?


As you have established that the prospect should buy a new point solution, now is the time to convince them that it is your solution that makes sense for them: it’s time for a new decision tree.


What you’ll need to do is build a new decision tree with a root node of “which next gen point solution to buy” and guide your prospects through these decision nodes, which will include different combinations of your ICPs’ key purchasing criteria (KPC).


KPCs should be as high level as ‘price’, ‘time to value’, ‘security’, ‘ease of use’, etc., and cannot be a long list of features included in one of your ‘basic’, ‘premium’ or ‘enterprise’ plans. Actually, it’s quite the opposite: if a KPC includes a feature listed with a ‘tick’ next to it under one of the plans on your website, it’s a clear sign that you’re in the weeds and need to zoom out; you need to bring the conversation to a higher level, to what adds value and actually matters. Again, you can’t expect your buyers to be ‘wowed’ by your file storage being 1GB per file or by 250 rather than 50 integrations on your basic plan, if you haven’t even clarified and agreed on what matters first.


Ok, let’s get back to decision nodes. As mentioned, in this case, decision nodes are different combinations of KPC based on how you and other players are positioned in the market. For instance, a decision node can be “low price, quick to value, and limited functionality”, whereas another decision node can be “high price, long implementation, and all the functionality an organisation will ever need”.


By building this decision tree with decision nodes matching your ICP’s unique KPC combination, you’re bringing the level of the conversation to what really matters, and simplifying the decision-making process. And by doing so, you not only establish trust and demonstrate understanding of your prospect’s priorities and needs, but also get alignment that you offer exactly what they are looking for.


At this point for your prospect, this isn’t another overwhelming conversation, because you’re feeding them with the key takeaways of an exhaustive market research rather than adding new items to their long list of things to google. After all, categorising a crowded universe of vendors and features into 2-3 combinations of KPC requires a deep understanding of the market. As Einstein famously said, “If you can't explain it simply, you don't understand it well enough.”


And if you do that on behalf of your prospect, choosing you will seem as obvious of a choice to your prospect as it does to you. If you find that it does not, then it's an indication that either that the particular prospect in question is not your ICP, or you need to revisit your market positioning.


So, good luck building your decision tree and testing your positioning!



Get in touch!

I’m always happy to discuss math theorems and how they are applied to the day-to-day business world! If you’d like to connect, you can find me on LinkedIn, or you can drop me an email at gc@oxx.vc.


Also, if you are intrigued to find out more about decision trees or market positioning, check out these links:



bottom of page