This post addresses a key question for start-ups, how do you model and forecast sales?
Please note that the technique below is best for enterprise-oriented companies rather than consumer Internet companies.
Forecasting Revenue
A key mistake start-ups make in raising money relates to how they model future revenues. This post explains a bottoms-up approach to forecasting revenue. My favorite bottoms-up forecasting method is the productive sales rep model.
In this model, future bookings are NOT a function of market share, size, and penetration rates ($500m market x .005 penetration, or $2.5m) but rather of how many mature sales reps are in the company and the expected sales rep quota and productivity.
A top-down approach is simply too hard to handicap and fails to ensure that a company matches an investment in sales resources with projected bookings and revenue.
First Model Bookings
Bookings = mature reps x quota per rep x productivity
Bookings = purchase orders
Mature reps = the number of reps with sufficient market and product experience to be effective (typically six months with the company)
Quota = bookings quota per year (typically $1-2m per rep in a start-up, and $2+m per rep in a mature company)
Productivity = percent of total quota achieved, on average, by the sales force
Therefore, for a start-up, with two mature reps entering the year, one rep joining in January, a $1.5m quota, and an expectation of 75% productivity, bookings would equal:
2.5 (mature reps) x $1.5m (quota) x 75% (average productivity as % of quota), or $2.8125m.
Then Assume Bookings Mix and Revenue Recognition Policy
To get to revenue, we then need to assume 1) a revenue recognition policy and 2) a bookings mix across license, maintenance, and professional services. This mix is typically 70% license, 15% maintenance, and 15% professional services.
Wrt revenue recognition, license revenue is recognized either at the time of sale for perpetual model or ratably for SaaS vendor, while maintenance and professional services revenues are recognized pro-rata over the course of the year/project. For example, assuming the bookings mix above a $1m purchase order (booking) on April 1 would contribute the following revenue in the year:
License revenues: $1m x 70%, or $700k
Maintenance revenues: $1m x 15% x 9/12, or $112.5k
Services revenues: $1m x 15% x 9/12, or $112.5k.
While there were $1m in bookings, revenues would be $925k, with the $75k difference on the balance sheet at year-end as deferred revenue.
The key issue is make sure that sales projections are tied to tangible investments in sales resources and are based on reasonable assumptions of sales rep productivity, time to maturity, and quota. Finally, think through how bookings translate to revenue. A sophisticated approach to the problem will go along way in gaining credibility with a prospective investor.
Very smart approach. How would you roll this together with Mark Leslie's Sales Learning Curve? Clearly there is a link.
ReplyDeleteThis is as good a method as any other, and lets face it, forecasting is pretty much about trying to take some control of the uncontrollable. Some of the things that might make this even more useful would be a common understanding of "mature salesperson", and "ability to defend pricing", particularly the upfront license. Perhaps the predictability of the model (in terms of variance) is more dependent on this quality of sales person, than any other part of the model. Thus, when you meet the venture, meeting the sales person and getting a handle on their quality is a key part of validating your revenue line.
ReplyDeleteWill - What do you mean by "The key issue is make sure that sales projections are tied to tangible investments in sales resources"?
ReplyDeleteThe major con of your approach is that if each new sales rep's marginal profits are positive, then a startup should hire a very large number of reps and forecast revenues accordingly without taking into consideration any market forces or law of diminishing returns.
Unless that's what you meant by the above phrase :)
Will,
ReplyDeleteI agree with you completely that the bottoms up forecast is likely to be far more accurate. But it sounds to me that having accurate numbers that don't predict a 10x return for a VC still doesn't get you anywhere.
There was an anecdote in "Founders at Work" about a former VC who tried to explain that his 5x return projection was a good investment because as a former VC he gave them the truth and din't inflate things to get the 10x prediction.
He finally had to change his model to a higher, bogus revenue number to get any interest in his company.
The fact of the matter is, with accurate revenue forecasting, very few companies would be VC investments.
Thoughts?
Thanks for all the good comments.
ReplyDeleteLet me first say that Mark Leslie cautions start-ups to hire slowly and once they know their sales life cycle. Direct sales is very expensive - you need to carry the fully burdened costs of the salespeople ahead of their productivity. Longer ramp periods, or sales head turnover extend that cost and further drain company resources. The key is to hire slowly and well - to test out key variables such as sales life cycle, ASP, pipeline coverage ratios, etc before hiring a bunch of reps.
For Cherif, I mean that sales are expensive and a headcount driven forecasting method explicitly ties revenue to sales costs. It helps sanity check the revenue production as every dollar of revenue comes with attendant sales costs -ie nothing is for free.
Finally, inflating numbers to secure funding is a bad idea. A good friend once told me that success= results -expectations. Something to keep in mind when putting together a plan
Will,
ReplyDeleteWhat have you seen in your SaaS startups for initial revenue forecasts. What has been the typical ramp for the 1-3rd year?
Hi Will,
ReplyDeleteNice post on an important topic.
One thing to keep in mind in forecasting revenue is that revenue recognition is complex and can result in massive revenue deferrals (e.g. Apple).
Many start-ups (and established companies) offer various concessions, including customer acceptance terms, custom development/futures, free upgrades, and free support/training to win business. This can have a dramatic impact on the start-up's ability to recognize revenue under accounting rules (SOP 97-2).
Start-ups need to keep this in mind when training their sales teams, negotiating customer contracts and forecasting revenue.
As a CPA & consultant focused on revenue recognition, I see this issue even with the largest companies in the valley.