Question and Answer: Using Analytic Business Intelligence to Drive Sales
The former head of Siebel's sales analytics team explains how companies can forecast sales much more accurately using effective analytics.
- By Linda L. Briggs
- October 21, 2009
"Sales forecasting has long been viewed as an art, and it is to some extent, but in my experience it is far more science than art," says Paul Staelin, co-founder of on-demand BI provider Birst and the former head of Siebel's sales analytics team. Sales forecasting is certainly critical. Forecast too high and operations will be saddled with excess inventory; forecast too low and operations won't be able to meet demand.
In this interview, Staelin draws on his years of experience to explain the three key phases companies move through in forming effective sales analytic solutions. Staelin points out the areas where many firms fall short.
BI This Week: You're an expert in sales analytics, having headed Siebel's sales analytics team before co-founding Birst four and a half years ago. How are sales analytics currently being used?
Paul Staelin: The sales organizations I've worked with over the years tend to evolve their analytics solution in three distinct phases. Although they may be pursued in any order, most companies follow the path below.
Phase 1 is sales performance. The first thing most sales organizations need to better understand and manage is their likely performance for the current quarter. This is important because by getting early insight into the likely or forecasted results (or both) for the current quarter, managers and executives can quickly spot problem areas and take action to address them -- instead of waiting until quarter-end and finding out they've missed their number. Combining forecast data from the CRM system with "actual" information from the back office systems and quota information from finance gives managers and executives the insights they need.
Phase 2 is sales effectiveness. Once the sales organization has armed managers and executives with the insight required to better predict and manage results in a given quarter, most organizations want to better understand how those results are delivered so they can improve them in the longer term.
To accomplish this, most companies begin to analyze and deliver what I call "effectiveness" or "velocity" metrics -- things that help the organization understand how effectively it turns opportunities into revenue. Key effectiveness metrics include average sales cycle, average deal size, average discount, average opportunity, close rate, average discount rate, average number of products per opportunity, and so forth. Changes in these metrics can significantly alter the organization's overall revenue generation.
Phase 3 is marketing effectiveness. Once most organizations get a handle on how they expect to perform this quarter, and have begun increasing their sales effectiveness, they typically begin to pay more attention to how effectively they're interacting with the marketing organization. This is an area that continues to be a source of contention in most organizations. Marketing becomes frustrated with a lack of lead follow-through by the sales organization, and sales becomes frustrated by the dearth of quality leads generated by marketing.
This divide is often best addressed by integrating lead, opportunity, and order data so both marketing and sales can see which leads turn into opportunities that close the most predictably, quickly, and profitably. For instance, it may turn out that the Widget2000 campaign in the Americas is generating leads that turn into opportunities 45 percent of the time and that 60 percent of those opportunities are closing. That may indicate that marketing should invest more in that campaign, and sales should ensure quick follow-up on any leads generated by that campaign.
Once this combined view of the marketing organization's effectiveness is shared across organizations, most companies are able to drive lead quality and value significantly, again enabling them to produce more bookings more quickly and more predictably from the same resources.
I'd like to drill down into Phase 1, where companies broadly measure how they're doing. How can a company's likely performance be measured and understood more effectively?
In general, by providing better insight into the expected and actual results of every sales representative, manager, and executive compared to quota, companies empower managers to quickly identify potential shortfalls. That allows them to more effectively take action while they still have time to influence the outcome. In essence, these organizations want better headlights, allowing them to see farther down the road and respond more quickly to sudden obstacles.
To most organizations, getting a solid handle on how the sales professionals in their organization will perform requires blending three types of information, often from three different systems: forecast (from the CRM system), quota (from the compensation system), and actual (from the back office system). The power of this comprehensive view of expected and actual performance is so compelling that showing this information for each of an example manager's subordinates was always the hit of every software demonstration I ever gave.
It's interesting that in most companies, the need for this kind of insight is often represented initially as a "forecasting" or a "pipeline analysis" problem. However, what they're really looking for is some way to better predict and manage results throughout their organization -- some way for them to know "Am I OK this quarter or not?" and to take remedial steps if required. Forecast/quota/actual analysis provides this insight more fully.
If a company is having trouble accurately pulling together those sorts of forecasts, what can they do?
It's interesting that forecast accuracy plays such a central role in most sales organizations' ability to deliver the desired results to the rest of the business, yet few organizations know how to forecast accurately. Forecast too high and operations will produce too much product and be saddled with excess inventory; forecast too low and operations won't be able to meet actual demand, putting revenue at risk and causing customer expectations to go unmet.
Forecasting has long been viewed as an art, and it is to some extent, but in my experience it is far more science than art. Take Siebel, for instance. While I was there, we deployed Siebel Sales Analytics internally. One of the most valuable things it enabled sales operations to do was to improve the forecast accuracy. Siebel sold large, enterprise software deals, with a few ultra-large deals disproportionately impacting results each quarter.
In this environment, Tom Siebel was only able to predict results for license revenue bookings at plus or minus ten percent or more in his mid-quarter calls with financial analysts. While this was more accurate than most enterprise software vendors of similar size, it wasn't accurate enough. Once Siebel Sales Analytics was deployed internally, the sales operations team was able to predict license software at plus or minus two percent per quarter.
The insight that enabled that radical increase in accuracy was a report that showed last quarter's pipeline with revenue closed by week compared to this quarter's pipeline with revenue closed by week. It turns out that despite what the money management firms say, past performance is indeed indicative of future results -- so the best way to predict how you will do this quarter is to compare the evolution of your pipeline this quarter compared to last quarter (or to last year, in a seasonal business).
Looking at Phase 2, in which you said that companies measure where the organization is most and least effective: Where are companies struggling there? What could they do better using analytics, and how?
Once companies have the ability to better predict and manage results every quarter, they usually turn their attention to increasing their effectiveness in delivering that revenue. This typically means analyzing their organization's ability to drive revenue quickly, predictably, and profitably. The metrics most organizations look at in this phase include average sales cycle, average deal size, average discount, and average number of products per deal.
These organizations are looking to find areas of their business where they can improve their effectiveness by spreading best practices more broadly, or by abandoning unattractive segments. For instance, opportunities with long sales cycles, small average deal sizes, and high discounts may be worth abandoning in favor of other areas where revenue can be delivered more quickly and profitably.
For example, if sales cycles with financial services firms in one part of the world are long, small and highly discounted, the organization may be better served by ignoring those opportunities in favor of manufacturing firms in a region that closes deals more quickly and with smaller discounts.
Phase 3 involves marketing and getting products into the pipeline. Where are the trouble spots?
In general, the hand-off between marketing and sales often is less effective than it could or should be. The disconnection between these two organizations is often driven by the way they are measured and compensated. Marketing is usually measured by how many leads are produced and sales is measured by how much revenue it produces. These two objectives are not equivalent and often lead to classic finger-pointing; marketing complains that sales isn't following up on the leads they produce, and sales complains that marketing is producing poor leads.
Even today, most companies find themselves in some variation of this situation. The best way out of this conundrum is to provide a unified view of the business to both organizations. In general, this means combining marketing data with both sales data and revenue data, then giving both sales and marketing a unified view into which campaigns are driving leads that turn into opportunities and then into orders.
Combining this visibility into a new measurement system in which marketing is compensated based on opportunities and revenue created can bring sales and marketing onto the same page. Marketing will invest more in campaigns that drive pipeline and revenue, and sales will follow up more proactively on the leads most likely to turn into opportunities and closed revenue.
What does Birst bring to what we've talked about here?
By combining the power of enterprise business intelligence with the speed and economics of SaaS delivery, Birst empowers more organizations to take the next step in managing their sales more effectively. As a result, we help more organizations understand and manage their businesses more effectively, more quickly, and more affordably than is possible through traditional solutions. Many of our customers can be up and running in a few weeks, so they can begin managing their sales organization more effectively right away. The monthly subscription means they pay for what they need when they need it. As the only self-service, software-as-as-service BI solution on the market, we make IT departments look good by helping them deliver high value more quickly and more affordably to their business users.