Most of us are familiar with Mark Leslie’s classic Sales Learning Curve and its implications for building the early salesforce at an enterprise startup. In short, it argues that too many startups put “the pedal to the metal” on sales hiring too early – before they have enough knowledge, process, and infrastructure in place – and end up with a pattern that looks like:
- Hire 1 salesrep, which seems to be working so we …
- Hire 2 more salesreps, which seems to be mostly working so we think “Eureka!” and we …
- Hire 10 more salesreps overnight
With the result that 8 of the 10 salesreps hired in phase three flame out within a year. You end up missing numbers and hiring a new VP of Sales who inherits a smoldering rubble of a salesforce which they must rebuild, nearly from scratch. The cost: $3-5M of wasted capital  and, more importantly, 12-18 months of lost time.
But let’s say you heed Leslie’s lessons and get through this phase. Once you’re up to 20-30 reps, you don’t just need sales to be working, you need to prove that you have attained the Holy Grail of startup sales: a repeatable sales process.
Everyone has their own definition of what “repeatable sales process” means and how to measure if you’ve attained it. Here are mine.
A repeatable sales process means:
- You hire salesreps with a standard hiring profile
- You give them a standard onboarding program
- You have standard support ratios (e.g., each rep gets 1/2 of a sales consultant, 1/3 of a sales development rep (SDR), and 1/6 of a sales manager)
- You have a standard patch (and a method for creating one) where the rep can be successful
- You have standard kit including tools such as collateral, presentations, demos, templates
- You have a standard sales methodology that includes how you define and execute the sales process
And, of course, it’s demonstrating some repeatable result. While many folks instinctively drift to “80% of salesreps at 100% (or more) of their quota” they forget a few things:
- The percentage should vary as function of business model: with a velocity model, monthly quotas, and a $25K ARR average sales price (ASP), it’s a lot more applicable than with an enterprise model, annual quotas, and a $300K ASP
- 80% at 100% means you beat plan even if no one overperforms  – and that hopefully rarely happens
- There is a difference between annual and quarterly performance, so while 80% at 100% might be reasonable in some cases on an annual basis, on a quarterly basis it might be more like 50%
- The reality of enterprise software is that performance is way more volatile than you might like it to be when you’re sitting in the board room
- When we’re looking at overall productivity we might look at the entire salesforce, but when we’re looking at repeatability we should look at recently hired cohorts. Does 80% of your third-year reps at quota tell you as much about repeatability – and the presumed performance of new hires – as 80% of your first-year reps cohort?
Long story short, in enterprise software, I’d say 80% of salesreps at 80% of quota is healthy, providing the company is making plan. I’d look at the most recent one-year and two-year cohorts more than the overall salesforce. Most importantly, to limit survivor bias, I’d look at the attrition rate on each cohort and hope for nothing more than 20%/year. What good is 80% at 80% of quota if 50% of the salesreps flamed out in the first year? Tools like my salesrep ramp chart help with this analysis.
But all that was just the warm-up for the big idea in this post: is repeatability enough? Turns out, the other day I was re-reading my favorite book on data governance, Non-Invasive Data Governance by Bob Seiner, and it reminded me of the Capability Maturity Model, from Carnegie Mellon’s Software Engineering Institute.
Here’s the picture that triggered my thinking:
Did you see it? Repeatable is level two in a five-level model. Here we are in sales and marketing striving to achieve what our engineering counterparts would call 40% of the way there. Doesn’t that explain a lot?
To think about what we should strive for, I’m going to switch models, to CMMI, which later replaced CMM. While it lacks a level called “repeatable” – which is what got me thinking about the whole topic – I think it’s a better model for thinking about sales .
Here’s a picture of CMMI:
I’d say that most of what I defined above as a repeatable sales process fits into the CMMI model as level 3, defined. What’s above that?
- Level 4, quantitively managed. While most salesforces are great about quantitative measurement of the result – tracking and potentially segmenting metrics like quota performance, average sales price, expansion rates, win rates – fewer actually track and measure the sales process . For example, time spent at each stage, activity monitoring by stage, conversion by stage, and leakage reason by stage. Better yet, why just track these variables when you can act on them? For example, put rules in place to take squatted opportunities from reps and give them to someone else , or create excess stage-aging reports that will be reviewed in management meetings.
- Level 5, optimizing. The idea here is that once the process is defined and managed (not just tracked) quantitatively, then we should be in a mode where we are constantly improving the process. To me, this means both analytics on the existing process as well as qualitative feedback and debate about how to make it better. That is, we are not only in continual improvement mode when it comes to sales execution, but also when it comes to sale process. We want to constantly strive to execute the process as best we can and also strive to improve the process. This, in my estimation, is both a matter of culture and focus. You need a culture that process- and process-improvement-oriented. You need to take the time – as it’s often very hard to do in sales – to focus not just on results, but on the process and how to constantly improve it.
To answer my own question: is repeatability enough? No, it’s not. It’s a great first step in the industrialization of your sales process, but it quickly then becomes the platform on which you start quantitative management and optimization.
So the new question should be not “is your sales process repeatable?” but “is it optimizing?” And never “optimized,” because you’re never done.
# # #
 Back when that used to be a lot of money
 You typically model a 20% cushion between quota and expected productivity.
 The nuance is that in CMM you could have a process that was repeatable without being (formally) defined. CMMI gets rid of this notion which, for whatever it’s worth, I think is pretty real in sales. That is, without any formal definition, certain motions get repeated informally and through word of mouth.
 With the notable exception of average sales cycle length, which just about everyone tracks – but this just looks at the whole process, end to end. (And some folks start it late, e.g., from-demo as opposed to from-acceptance.)
 Where squatting means accepting an opportunity but not working on it, either at all or sufficiently to keep it moving.