Last week I posted about how unhelpful pipeline coverage metrics are (among other things). I said you might as well look at sunspots to predict what you will sell. My business partner challenged that. He thought that reading tea leaves was more appropriate. I don‘t know much about that. Here’s what I know about sunspots.
Sunspots are temporary darker (and cooler) areas on the sun's surface. Their first recorded observation dates to before 800 BC.
In 1878, economist William Stanley Jevons speculated that business cycles are affected by sunspot activity. It's not as far fetched as it may seem. (Kind of like pipeline coverage metrics. More on that shortly.) His theory was that sunspots might cause variations in weather and thus agricultural output.
Since then, many have tried—with little success—to uncover this relationship. Lack of evidence has not stopped people from trying. But lack of evidence is hard to shake. The theory is now widely discredited. This has led economists David Cass and Karl Shell to more recently coin the term sunspots to mean a random variable that has no relation to a measured outcome. Hence my musing about sunspots and pipeline coverage.
Pipeline coverage ratio or “PCR” (total pipeline with a close date within a period, divided by a goal for the period) is a time-honored metric to gauge if you will meet your goal.
Does PCR do this? We measured it. The results are disappointing because:
PCR varies widely from quarter-to-quarter. This makes it hard to take seriously any notion of a minimum required ratio—unless you set some upper limit which makes the ratio so large that it will be discouraging and unhelpful.
PCR obscures the underlying causes of a potential shortfall. It mixes the effects of deal sizes, win rates, deal stages (pipeline quality), and new opportunity generation. PCR may tell you if you need more pipeline. But not how much more. And you don’t need a metric to tell you that.
In defense of conventional wisdom, it seems logical that the bigger your pipeline the more you will sell. But the empirical evidence doesn't support that. There are so many other factors that affect sales: deal concentration in your pipeline, distribution of sales stages, new pipeline generation. And maybe sunspots. Or tea leaves.
Here is what that looks like in practice. The two plots below show historic 90-day PCR values (filtered by new and expansion business, for sales qualified and above deals, with close dates in the next 90 days) relative to actual amounts won over the next 90 days from each date. In other words, instead of dividing by a revenue goal, the actual amount won over the next 90 days is used. Call this a 90-Day "sliding-window" PCR. The gray vertical bars indicate the first week of each quarter. If PCR were a good predictor, each chart would be a flat line. Instead, we see a lot of variation. It ranges from 3:13 for company A, and 7:16 for company B. Conventional wisdom is that you need 3x your goal. Although when people quote that ratio, they rarely mention when this is measured. And that matters a lot. I think that they mean as measured shortly after the start of the quarter.
Looking only at the start of quarters is similarly unhelpful. For company A this ratio ranges from 4 to 10x. Company B, from 7 to 15x. So much for needing 3x.
More important than the actual ratio is whether pipeline coverage is predictive. For these companies, if you used this ratio to predict what you will sell, you could be off by more than 2x. Maybe better if you picked a ratio in the middle of the range.
Some of you are thinking, “We revise our PCR window throughout the quarter.” On day one you look at pipeline with a 90-day window for close dates. On day two, an 89-day window... Each day, you also update the remaining goal based on closed sales to date and calculate PCR for that new goal. Call this the Intra-Quarter PCR.
We measured that too. (Again replacing the goal with the actual sales closed in the quarter.) Here’s one example (company C, admittedly one of the worst offenders we studied). Like the sliding-window examples above, the starting ratios are quite different in each quarter—ranging from 5 to 14x. In the first month, the actual PCRs for this business were relatively stable. You could use the month 1 pipeline to predict sales—if only you knew whether to divide by 5 or 14. After day 45, the metric is wildly unpredictable.
For comparison, here’s a chart of sunspot activity. That looks more predictable than PCR to me. Of course, it is not predicting your sales or anything else. But the cycle is ok at predicting the next peak. They follow 11-year cycles (on average). Except when they follow nine- or ten-year cycles.
This regularity has led researchers to try to correlate economic results with sunspots. Here’s an example in what appears to be a peer-reviewed paper. At the time of its March 2023 publishing, when the federal funds rate was 4.57%, the author was (using sunspots) predicting “in 2023 this rate will increase to 1.996%.” (Such precision and authority!) The rate eventually hit 5.3%. In another paper (probably not peer reviewed), the author seems to see correlations to recessions that aren’t there in my eyes. So much for sunspots predicting economic factors.
Was Jevons's 1878 sunspot theory of economic cycles more or less useful than using PCR to predict sales? In retrospect, I may have short-changed PCR a bit. I think it slightly edges out sunspots. But, Is PCR useful? Does it tell you how much more pipeline you need? Which deals to work on and when? Is it better than reading tea leaves?
You be judge.