Imagery Majestic - Fotolia
As the head of sales, you're in the hot seat several times a day. But this time feels different.
You have only a week left until the quarter closes, and you're meeting with your CEO to review the sales pipeline. The CEO asks a seemingly simple question: "Which deals will close?" But the question is lost on you, because you have your own questions:
- Which deals will really close this quarter?
- Which deals should the sales reps focus on?
- What can reps do to ensure that those deals will close?
- Can the sales team even trust the forecast?
It's not a good place to be. If you're still having these questions at the end of the quarter, it may be too late to change things. Sadly, you're not alone. Sales funnel management is a challenge for myriad companies, but it's essential to closing more deals.
As B2B buyers take more control of the buying process and wait to engage with sales reps, Gartner said that companies are seeing more unpredictable sales funnels (the process by which a prospect becomes a buying customer) with wider tops (i.e., more new leads from, say, content marketing) and narrower, more unpredictable bottoms (closed deals) with lower conversion rates. The challenge: How can sales and marketing develop a more predictable sales funnel so you can increase revenue in a more reliable way?
While it's not a silver bullet, that is the promise of predictive analytics for sales funnel management. From the beginning of the quarter to the very end, predictive analytics helps an organization better understand how to improve results by focusing efforts on the best potential business and on turning around likely prospects that went dark somewhere during the process.
This is one reason why, according to a recent Salesforce survey of 2,300 global sales leaders, 33% of high-performing sales teams have deployed predictive sales analytics -- a rate of four times that of underperforming teams. "For these leading teams, analytics likely provide visibility into accounts and helps dictate where to focus energy for the most productive customer and prospect conversations," Salesforce reported. The survey also found that 58% of respondents expect to increase their use of analytics by year's end.
Better models for better results
For years, companies have modeled composites of the "ideal" customer, the customer most likely to close and spend the most on a given product. The problem is that those models are limiting, and can encompass only a certain number of variables. Such models are often manually created and updated and are based heavily on qualitative observations rather than quantitative data.
By contrast, predictive analytics joins "first-party" data from internal systems such as CRM and marketing automation platforms with external third-party data from hundreds or thousands of data sources. The goal is to help the organization make better decisions at every stage of the sales funnel, such as the following:
- Which new leads are most likely to close in the least amount of time and spend the most money?
- Which opportunities are most likely to close in the least amount of time and at the highest deal size?
- Which customers are most likely to buy additional products? Which are most likely to defect?
- What will my pipeline be?
More data, better analytics, better decisions and better actions
Eric EsfahanianGryphon Networks Corp.
These analytics are based on predictive models that correlate past performance against hundreds or thousands of disparate signals, such as new office openings, recent interactions with your firm and prospect company size. Vendors have various predictive analytics offerings that automatically generate models to predict future performance so you can drive better results. Unlike historical reporting, predictive analytics are based on current data that is updated nightly or in real time.
Vendors often provide black-box models -- whose algorithms are opaque and cannot be adjusted by users on their own -- and model building. They typically use the following types of data:
- Firmographics. "Demographics" about the company, such as size, location, revenue and more.
- Technographics. Technologies in use by the prospect or customer.
- Intent data. Data coming from social media, activity on the Web (e.g., "What kinds of white papers has a prospect downloaded?"). This describes what a prospect likely intends. For example, someone who downloads 10 white papers about accounting systems may plan to acquire an accounting system.
- Engagement data. Level of engagement with your company. For example, how often has a given lead downloaded white papers from your website?
- Sales rep activity. The actions the sales rep has taken. The predictive analytics system may compare the activity of the best reps against all others in order to predict a rep's pipeline, close rate and more.
Predictive sales analytics across the customer lifecycle
According to Gartner, the following analytics predict what will happen across the entire customer lifecycle, from new lead to close to customer renewal, upsell, cross-sell and defection:
- Prospecting: Identifies the best companies and people to prospect.
- Lead scoring: Helps reps prioritize the leads most likely to turn into closed deals, based on buying signals.
- Opportunity scoring: Provides a continually updated opportunity score based on deal size and likelihood to close.
- Pricing. Helps optimize pricing for improved margins.
- Customer renewal and up- and cross-sell. Predicts customers most likely to renew, purchase more products and even defect.
Better leads, more deals
Predictive analytics have improved sales efficiency and driven revenue growth -- both important for better sales funnel management. For example, B2B companies that actively use analytics gain 15% shorter average sales cycles and 11% larger average deal size, according to Aberdeen Group.
Sales analytics company InsightSquared experienced just that after bringing in a marketing and sales platform to get better intelligence on leads. "We were able to suppress a significant portion of our lead flow that was low value," Adam von Reyn, director of growth marketing, said. "The team could spend more time working with high-value leads -- which convert three times as highly as the rest."
Blue Jeans Network, a videoconferencing provider, targeted a new batch of prospects that were already in its database, but sitting idle. By adding external data, vendor 6sense identified 1,800 prospects in the system that were ready to buy, but weren't being actively pursued by the sales team. The 6sense model converted these leads at a rate that was 20 times higher than leads scored by Blue Jeans' internal system -- and with 66% fewer touches than with Blue Jeans-scored leads, according to 6sense.
Understanding the predictive analytics landscape
Prescriptive analytics: It's essential to go beyond predictive to do "prescriptive analytics," determining the optimal next step based on analytics, Josh Evans, senior vice president of sales at sales acceleration vendor Velocify, said. "Prescriptive analytics enables our sales reps to normalize their individual data patterns to create actionable next steps," he said.
Similarly, vendor Lattice Engines recently added information to help sales reps in the context of their work, telling reps the underlying attributes that drive predictive lead scores, products the prospect is most likely to buy and talking points for the prospect. "Companies that gain the most value from predictive analytics ensure that there is a strong process for delivering leads to sales and sales development in a way they can and will consume effectively," Kerry Cunningham, research director at SiriusDecisions, said.
Account-based marketing (ABM): ABM identifies the best accounts and coordinates marketing and sales activities around those accounts. Companies are beginning to use predictive analytics to identify the best target accounts, create the right messages and tactics to convert accounts, and identify the accounts most likely to buy more -- or to churn. InsightSquared's von Reyn said, "Our predictive account score helps us understand which accounts are most likely to convert into opportunities and deals."
According to Megan Heuer of SiriusDecisions, predictive analytics enables companies to scale their ABM efforts, which can be a difficult task as companies expand.
Packaged offerings: Unlike many other technologies, sophisticated predictive analytics can be challenging to do on your own. They require vast amounts of raw data; data scientists; massive computing horsepower; integrations to internal CRM, marketing and other internal systems; and thousands of data elements. That's why more and more companies are turning to packaged solutions, often provided by software as a service vendors. InsightSquared tried to build in-house, but "to get the results we desired, we needed to go to the experts," von Reyn said.
Getting ROI from predictive analytics
Although many predictive analytics offerings have similar underpinnings, differences remain. They may take days to several weeks to deploy, depending on how productized (prebuilt, out-of-the-box) their offerings are. Vendors offer varied integrations and data sources, so you should make sure they have the ones that you need. And some predictive models employ machine learning to continuously improve over time.
It's quite possible that you will deploy a vendor's offering because of a lack of data science expertise at your company. Be aware that most predictive analytics models are black box. Most models require a significant number of won deals in order to be useful. Once the vendor has constructed its model based on your data, you should test that model against your set-aside historical data, as well as new data, as Blue Jeans Network did. You should also plan to test multiple vendors' models so you can choose the most accurate one. As always, check references to understand other customers' deployment and support experiences with the vendor.
Keeping it simple is good policy. "It can get complicated pretty quickly," said Eric Esfahanian, senior vice president and general manager at Gryphon Networks Corp. "For sales managers, simplicity is king. We aren't looking to become junior data scientists, nor are we looking to hire one to properly process the information we need. Over time, you will have a mountain of raw data, so the challenge ... is to use a view that presents only the insights that are most meaningful to your business."
Marketers, time to get on board with your sales teams
New tools to help sales teams close the deal
Is the sales funnel dead?
Why marketing-only reporting software is dying