“My forecasts are always 100% accurate. Unfortunately, my customers don’t always buy when they are supposed to.”
BEST cash advance— Anonymous Salesperson
While customers don’t always do what we expect of them, sales pipelines do behave in a logical manner — as long as you know what to look for.
If your company’s sales process involves multiple stages where salespeople interact with numerous decision makers, knowing which data to extract from your CRM system to build a forecast can be difficult — particularly when many CRM systems don’t automatically track what’s most valuable. Using these five pipeline metrics, you can significantly improve your forecast accuracy:
- Opportunity Quantity — the simplest of the four, this is a measure of how many opportunities are in the pipeline
- Opportunity Size — another easy one, this is the projected dollar (Euro, Yen, etc.) value of each opportunity and is often expressed as an “average deal size”
- Conversion Rate — assessed on a stage-by-stage basis, this measures what percentage of opportunities at one stage typically make it through to the next
- Velocity — also measured on a stage-by-stage basis, this is a measure of how much time opportunities typically spend in each stage (Note: this one is often left out by CRM systems – if so, search for a plug-in that will help you track this metric)
- Opportunity Quality — the hardest one to quantify, this is an assessment of how precisely an opportunity matches a predefined definition of a well-qualified sales lead (another one often left out by CRM systems, this can be easily added using a custom field)
Turning our attention to how each of these metrics will enter our forecast model, Opportunity Number and Opportunity Size are fairly straightforward. Multiply the former by the latter (assuming the latter is expressed as the average deal size), and you have the overall value of your pipeline if everything closes (don’t we wish!). This will act as the baseline upon which the rest of our calculations will be based.
The next metric, Conversion Rate, will be used to peel off the value of those opportunities in each stage that, based on historical analysis, will not close. If you’ve done your homework and understand the conversion rates for each stage, you can determine how much revenue you can anticipate coming out of each stage by multiplying the value of opportunities in each stage by the conversion rate for that stage and by the conversion rates for all subsequent stages. An example will help illustrate the math required here…
- Suppose we have four sales stages with the following conversion rates:
- Between stage one and two: 25%
- Between stage two and three: 35%
- Between stage three and four: 50%
- Between stage four and close: 65%
- To figure out the value of your pipeline qualified by conversion rate, you would add together the following values:
- (The value of opportunities in stage one) * .25 * .35 * .50 * .65
- (The value of opportunities in stage two) * .35 * .50 * .65
- (The value of opportunities in stage three) * .50 * .65
- (The value of opportunities in stage four) * .65
Velocity enters into our model to allow for sales cycles that are longer than the desired forecast period. If, using historical data, we know the average time-in-stage for each stage, we can calculate a forecast for a given time period by taking into account the velocity of our pipeline. Here’s how: if our above example has a one year sales cycle that’s perfectly uniform (i.e., opportunities are in each stage for three months), a quarterly forecast would simply be the value of opportunities in stage 4 multiplied by the corresponding conversion rate of .65.
The last variable to consider is Deal Quality. This is tricky because how well an opportunity adheres to our definition of a top-tier prospect is subjective. As such, deal quality is usually represented by a weighting factor applied to each opportunity. For example, “hot” opportunities (i.e., those that match our top-tier prospect definition perfectly) may have a weighting factor of 1.0 – meaning that 100% of their value is passed along to the formula we’ve developed above. Prospects that are only “warm” (i.e., those that loosely match our top-tier prospect definition) may have a weighting factor of .65 – meaning that only 65% of their value will be used in our calculations. While they introduce a little “slop” into our math, these weighting factors give your salespeople a badly needed means to qualify opportunities in their individual pipelines. This can go a long way toward neutralizing an accurate forecast’s worst enemy: sandbagging. What’s important is to have a clear top-tier prospect definition that everybody agrees upon and a clear understanding of the weighting factors associated with the deal quality labels you allow your salespeople to use (e.g., hot/warm/cold).
While this process can be refined in a variety of ways, it is an excellent starting point that you can use to develop your own forecasting model. What’s critical is to view the development of your forecasting model as a process and not an event. Conversion rates and velocities will vary with the ups and downs of the economy, the season, and the competitive landscape. As such, you will need to constantly refine your forecasting formula based on how these and other variables impact your key pipeline metrics over time. In this regard, accurate forecasting is an historical science – the more you look to the past, the better you will become at predicting the future.
Update: In a continuation of this discussion, we have a new blog post that describes the use of baselines and trends to developing sales forecasts. Entitled “Using Baselines and Trends to Refine your Forecasting Process,” this post discusses how to use historical data to improve forecast accuracy.
Easier said then done.. With all due respect to my colleague and writer of this brilliant piece. In my 7+ years in Sales Ops I have found that many times sales will call a number that they will close (Forecast) but they will not commit specific deals. As per them they are not sure if deal X will close or not. it is likely that it will close but what if it doesn’t then they will look bad.
There is also a concept of ‘rolling garbage’ in pipeline. If the deal didn’t close this quarter sales might simply change the close date and it will become a key deal for next quarter. This deal is never meant to close rather this is a place holder for number. The close date will keep changing for this deal.
Sales Ops person needs to do some investigative work as well and nail down these rolling optys so as to mark them lost in the system and force the sales team to create fresh and new leads.
Hasan Qureshi[Click to quote this in your comment]
Hasan,
You raise excellent points. Forecast accuracy is an elusive goal that must be tackled from multiple angles. While intelligently extracting and using data from your CRM system is a great starting point, improvement can only come with ongoing examination of what causes inaccuracy to creep into the picture. Sandbagging and “rolling garbage” are definite culprits in this regard.
Your suggested solution, investigation on the part of sales operations professionals overseeing the forecasting process, is spot on. This is why we’ve always considered good pipeline reports that are understood and used by first-line sales managers (FLSMs) to be a vital part of successful CRM implementations. A pipeline review process that is a standard part of how sales teams operate not only raises team member accountability, it also provides an easy way to spot irregularities over time.
When opportunities never die (the “rolling garbage” problem you point out) or sales reps engage in sandbagging, the velocity metric will suffer. Looking for opportunities that have a long “time-in-stage” will reveal the former, while a velocity metric that always goes down near month or quarter end is a clear sign of the latter. It still requires a diligent sales operations professional to sniff out these problems, but regular reports that are read and used by FLSMs can provide an excellent way to engage them in resolving these issues when they are identified.
- Tom
Thomas Barrieau[Click to quote this in your comment]
Quality is in the eye of the beholder and most sales reps believe their next sale will be the greatest. That’s good too, you want sales people that are excited. But I’ve never measured velocity. Reading through your post I think it has some merit but in the end the client has to want to sign. Speeding up the process may only hurt the long-term aspect of the relationship.
Levi Spires[Click to quote this in your comment]
Levi,
Thanks for the comment. I agree that “quality is in the eye of the beholder” and that a sales rep’s optimism can mask an opportunity’s true likelihood of closing. There are two ways of addressing this…
One method would be to hold salespeople accountable by reporting back to them how well they are rating the quality of their opportunities. If they are consistently optimistic, they can be coached to be more realistic. Likewise if they are consistently pessimistic. This approach has the virtue of being a simple and quick solution. It’s downside is that management intervention (i.e., coaching) is required.
The other alternative eliminates the requirement to do reporting and coaching, but it is more complex. This involves using the CRM system to calculate each rep’s historical performance as far as how they rank their opportunities. A coefficient can then be developed that adjusts for individual optimism/pessimism. This latter approach only makes sense if you’ve got a relatively stable workforce and enough historical data to compute meaningful coefficients.
Regarding velocity… Your point about clients having the ultimate say on when they buy is well taken — I’ve never been a fan of artificially rushing the sales process. The point of using velocity as a pipeline metric is that a given company’s sales process will have an average velocity. Factoring that into the computation of a forecast is, therefore, always going to improve your accuracy, particularly if you take into account variability that is due to seasonality or product line. While there will always be end-of-quarter attempts to pull in deals that are close to closing, I see velocity as something that should never be manipulated as part of the forecasting process; it’s merely one of those variables that must be factored in to improve accuracy.
Any attempt at improving pipeline velocity should be done independently of the forecasting process, as part of a systematic effort to improve the efficiency of the sales process. While this is a good goal (IDC research indicates that most IT customers wish that purchase cycles were shorter), it is not something that can be done quickly and must be done in a manner that does not damage customer relationships – just as you point out. If efforts to shorten the sales cycle are successful, the velocity variable in forecast calculation can then be adjusted accordingly.
Thomas Barrieau[Click to quote this in your comment]