Usage Based Pricing — A Quick Primer from a Strategic Finance Lens
We recently concluded our one year anniversary of what is colloquially referred to as usage based pricing, which I’ll reference henceforth as consumption based pricing (“CBP”) at Fivetran. Since embarking on our journey a year ago, I’ve noticed a number of companies question whether they should make the move from subscription pricing to CBP. Part of this interest stems from the paradigm shift of readily available cloud warehouses that enable companies of any size to be data driven organizations. The other interest likely comes from the wild financial success companies with CBP models (e.g. Datadog, Snowflake, Twilio) have achieved all whilst catering to customers of all shapes and sizes.
Despite the tailwinds favoring CBP, relatively little information can be gleaned from these example companies’ websites, so much of our current infrastructure and standard operating procedures were therefore derived from trial-and-error (and many changes to the financial model, countless enablement sessions, and a patient, world-class sales and customer success team that would be the envy of most management teams), so without further ado, here are some thoughts, takeaways, and suggestions for others considering jumping off the deep end.
I will preface by noting that this is written from my strategic finance vantage point. I tend to think about the structure of companies in the following manner: If I perform well financially, my enterprise value and stock price goes up. If my stock price goes up, I’m able to retain and attract the best talent. If I’m able to retain and attract the best talent, they will build a fantastic product and customer experience. If I have a fantastic product and customer experience, then I will perform well financially. Therefore, my biggest concern should be performing well from a top-line perspective. But my colleagues were quick to point out that ease of deployment, time to deployment, customer satisfaction, minimal customer tension are paramount to anything else — so a special thank you to my colleagues and others who debated this topic during this deployment. (I’m also aware that the cycle I’ve noted above closely resembles the Chicken-and-Egg Paradox). Thanks Mona P for the edits and feedback.
Overall Thoughts:
1. Is CBP right for your business? This should be the first step for you to tackle, and I think the appropriate way to answer this is understanding the market itself. In the event your business is a greenfield opportunity, then you have the flexibility to do as you wish — for the rest of us that are playing in brownfield opportunities and competing against incumbents and other competitors, do our competitors employ CBP? If the answer is yes, then I think it’s appropriate to embark on the CBP journey. After all, your customers will host bake-offs and it’ll be appropriate to compare similar pricing schemes. If the answer is no, then you’ve reached your last step, which I refer to as a “Come to Jesus” moment. Find your harshest critic, and ask them if your product is differentiated enough to warrant a different pricing structure. If the answer is no, you risk alienating customers and prospects that wish to evaluate similar pricing schemes vis a vis other competitors.
2. Is CBP pricing right for your customers (I.e. are you doing right by your customer by charging them for only using the product)? The answer here should be a resounding yes for almost all companies. After all, nobody wants to be guilty of selling shelf-ware to their customers. But still, there could be edge cases where CBP isn’t a very effective pricing strategy. For example, companies with a seat-based license model that deploy technology that is difficult to measure consumption patterns likely would be better off keeping the status quo SaaS model. Which brings me to a set of takeaways.
Takeaways:
1. How exactly should I price for CBP? This is THE most critical part of the journey and one that you need to nail shut, seal closely etc. etc. Put simply, your medium of exchange needs to be airtight. How exactly do you get there? There are really four key elements that a good credit/medium of exchange needs to have IMO a) it needs to be auditable, b) it needs to scale as your customers scale, c) it needs to be logarithmic, and d) it needs to be difficult to debate with minimum volatility (SUPER IMPORTANT!) The medium of exchange needs to be auditable because you want to be able to identify any ‘leakages’ and scenarios where your customer was unduly charged so you can rectify it prior to the charge being made. It needs to scale as your customer scales because you want to be able to charge your small customers tiny amounts so that both your growth (top-line growth) and their growth (customer success) is pari passu. It needs to be logarithmic, not linear, because we as humans experience the benefits of pretty much everything in a logarithmic manner (I love this video by Hannah Fry that gives a quick primer on logarithms in case you need a calculus refresher), software is no different. And then lastly, but most importantly, you want your medium of exchange to be obscure enough that your customers aren’t debating usage and gawking at it with an eagle eye. I love the Snowflake “compute time” charge. It’s tangible enough that you’re aware you’re using it more, but obscure enough that no one is too microscopic with it.
2. How can I test this? We eventually settled on a ‘monthly active row’ calculation for charges. This ended up being a good exchange medium since the sales pitch was akin to ‘LITERALLY pay for what you use’. However, we weren’t really sure a) which customers would pay more or less than what they were currently paying thus b) how this pricing change would impact our revenue and c) what the user feedback would be. The latter issue was the easier of three — we could deliver user surveys to get an understanding of how our customers would welcome the change. Overwhelmingly, the feedback we got was ‘everything else in the stack is charging this way — makes sense’. The revenue part was significantly trickier since we didn’t have any past data on this metric. Our best way to tackle this was by creating a financial model, using empirical data on a similar metric, adding a few assumptions, and then using the output to estimate which customers we would lose revenue from. We then counted up which customers we would gain revenue from to get an understanding of how we could come out breakeven and transition as many customers as possible.
3. How should customers consume credits? This one will likely lead to the most heated debates amongst team members since each stakeholder approaches this question from a different perspective, but ultimately my position on this question is you want to a) minimize volatility in your company’s revenue, after all this protects your ability to be a going concern in perpetuity b) give your customer predictability, after all, nobody likes volatility in their bills and c) anticipate growth in your customers before they understand their growth so you can be proactive about “upsell” discussions. The best way to have customers consume their credits is via a monthly minimum commitment for customers on an annual deal and a pay-as-you-go for monthly, non-commitment customers. The more a customer commits, the more they should be able to tap into additional discounts.
Suggestions:
1. How should I apply discounts? The tough part of consumption based pricing for sales teams is the inherent nature of how it’s deployed. It’s more often than not tied to a transparent pricing structure that’s publicly available (e.g. a standard credit is $xx, an enterprise credit is $xx). This leads the sales team to win more on a) superior product differentiators b) command of the message and c) superior customer service and fantastic SLAs. Still, let’s not forget that many companies have whole procurement teams dedicated to lowering the price on a purchase order. Therefore, you have to ask yourself — what exactly would I like in exchange for a lower price? Having a customer commit to a monthly spend (the higher the spend, the higher discount) creates a fine incentive. You, as the company, get to have a stronger grasp of your forecasted revenue and what expenses you can associate to them. The customer also gets predictability in their monthly spend and credit price.
2. Roles and responsibilities. This one is tricky to get right since it really boils down to your team, their strengths, and the sales cycle of a customer. Still, I’ve reached the conclusion that the best way to achieve velocity of lands and stickiness of usage (land and expand) is by deploying the tried and true AE/CSM model where the AE is obsessed with courting new customers (throwing over the fence) and a CSM is obsessed with onboarding new customers and deploying the expansion playbook. We toyed with having a hybrid model, but ultimately concluded that the skill sets are vastly different, the win rates are different, and therefore you’re asking someone to change speeds and deploy different patience levels in a day and it’s enough to give a quality sales rep whiplash.
3. Getting compensation right. Because a customer’s revenue is unknown during the first few months, you’re left with an unknown way to comp AEs/CSMs. There are a few different strategies, but I’ve concluded the following are fair, gives each party something to look forward to, and compensates the right behavior. :
- AE Compensation: commissions are given for annual deal bookings with a “true up” at the conclusion of the 12-month period. The true up should be calculated by averaging usage over the duration of the 12 months. If average usage > ACV Booking then a sales rep should get additional credit. I won’t dive into ACV > average usage and whether or not it warrants a ‘claw back’, since I generally see claw backs as not worth the hassle — you’re likely to create a ton of friction in order to save a nickel here or there.
- CSM Compensation: commissions are given for net retention and logo retention with a quarterly reset on an AM’s book of business. Obviously, net retention will be a key financial metric for the health of the company, and there’s few individuals that have a greater influence on this metric than your CSM org. Still, you really want to be obsessed with customers ramp-up, their stickiness and loyalty to the product, so having a good logo retention metric helps your compensation model be ‘forward-looking’.
4. Avoid the rollover. My viewpoint on this matter mostly from an accounting, revenue recognition standpoint (I forewarned ya!). But, there’s a good argument for removing rollover from your initial sell and during your renewal period. If you let customers purchase an annual deal, but note that there is a rollover clause, and then follow up with a wink and a head nod, it’s very easy for the buyer to bulk up front — after all, the credits rollover. Unfortunately, this creates a mismatch between your commission model and your top-line revenue. Why? If the deal ends up with a duration of 20–24 months, you all of a sudden have a sales and marketing expense that is sky high relative to revenue. Moreover, having a rollover clause makes it impossible to understand when a customer is ‘fully ramped’ (fully ramped here meaning, at what point is a customer spending what they actually committed upfront?). Removing rollover allows you to track these metrics much more effectively. Besides, your CPAs will thank you (after all, who wants to be on the bad side of an auditor).
