2. As companies start experiencing the benefits of fit-based predictive scores, they typically seek out more
places where predictive can add value. One logical step is to add predictive behavior scoring. Infer mines
the full spectrum of activity data being collected by your marketing automation platform including every
email click, website visit, etc. Machine learning is then used to weigh each signal appropriately in order to
predict the likelihood of an outcome (e.g. conversion) within a set time period (e.g. next three weeks).
This predictive playbook will describe how companies can supplement fit-based predictive lead scoring
with an Infer behavioral scoring model. There are several ways to operationalize this -- from advanced sales
prioritization and activity-based triggers, to measuring account-based marketing engagement and
campaign effectiveness. For the purposes of this playbook, we’ll focus on the winning strategy of
resurfacing gold out of nurture by identifying previously disqualified prospects that have re-engaged and
are worth another look.
Many Infer customers enjoy a healthy inbound lead flow that fuels sales and marketing, and use Infer’s
fit-based predictive scores to easily qualify which leads are the most well-matched for the company’s
products. Infer’s custom-built fit models put inbound leads into categories from Infer A-Leads (best fit) to
Infer D-Leads (worst fit), and can automatically route all A and B-Leads to the sales queue. Infer C and
D-Leads – along with the good leads that are disqualified by sales – are typically moved into a nurture
database for marketing follow up. But as these archived leads accumulate, it becomes increasingly difficult
for the marketing team to monitor all of the older leads and find accounts that have re-engaged and are
likely to buy.
Introduction
Getting Started
3. Once that process is demonstrating results, most companies ask Infer to build a second predictive model,
this time pulling in key activity signals from a marketing automation platform like Marketo. Unlike traditional
marketing automation lead scoring that requires users to manually add points for given actions, Infer’s
behavior scoring model uses powerful machine learning to mine the full spectrum of a lead’s activity data –
including concentration (or decay) of activity and breadth of engagement – to predict which prospects will
convert in the next three weeks.
Prioritizing “Older” Leads – True
Nurture vs. Neglect
By closely observing behavior scores to find key moments in a lead’s buying cycle, you can pinpoint
the right time(s) to reach out again and trigger the appropriate actions therefore empowering sales
to own the moment in a first-conversation or second-chance engagement.
Website
Sales
Queue
Good Fit
Leads
Bad Fit
Leads
Qualified
Archived
Marketing
Nurture
Database
High
Behavior
Score
4. Closing New Deals from Nurture
Once a new behavioral model is in place, you should start monitoring these predictive scores and establish
a manageable way to work high-scoring leads from the nurture bucket. Some marketing teams start by
placing unqualified leads into an unassigned “Up-for-Grabs” queue where sales development reps can find
additional leads to work when they have time. Leads taken from this queue can be easily prioritized based
on their behavior, with the highest scores indicating that the contact or account is showing fresh activity
indicative of purchase intent. This also gives reps increased confidence during outreach because the scores
are driven entirely by a prospect’s interaction with your company.
This approach brings the best nurture leads back to sales without fully opening the floodgates to
thousands of archived leads. And as more and more reps start working these leads, the team’s confidence
in both the fit and behavior scores grows because it becomes easy to see how much better higher-scoring
leads convert. If you break leads into behavior buckets, from Infer 1-Leads (those showing the most buying
behavior) to Infer 4-Leads (least buying behavior), it might play out something like this:
By renewing sales’ focus on select older leads, one Infer customer closed nearly a half-million dollars
in deals with prospects that otherwise might have languished in nurture.
Behavioral Score
Buckets
Leads Grabbed from
“Up-for-Grabs” Queue
Converted Opps
Lead to
Opp Rate
Conv. to
Opp Rate
Infer 1-Leads
(scored 70+)
Infer 2-Leads
Infer 3-Leads
Infer 4-Leads
407
335
269
682
115
94
72
240
53
30
18
40
13.02%
8.96%
6.69%
5.87%
46.09%
31.91%
25.00%
16.67%
5. While predictive scoring is most often used to prioritize new leads, there are clear benefits to pulling in
behavioral data and monitoring aging leads over time. Several Infer customers have proven that predictive
can deliver a material return on investment when scores are operationalized effectively across sales and
marketing teams. Want to learn more about how your business can adopt predictive to drive success?
Get in touch with Infer today.
When looking at the full spectrum of your leads, it’s helpful to combine your fit and behavior scoring
buckets into a simple 4x4 matrix so that you can optimize programs around various segments. For example,
conversion multipliers across different lead classifications often look similar to this:
There are many ways that companies are leveraging this insight to increase sales and marketing
productivity. For example, by sending Infer A1-Leads (which represent a disproportionate percentage of
your overall conversions) directly to your best reps, you can close those key deals faster. Leads that typically
represent a lower return but are showing a high level of buying behavior, like Infer C1, D1, C2 and C3-Leads,
can be worth investigating if your team has the bandwidth – perhaps using an “up-for-grabs” queue as
described above.
In addition, companies use this grid to create service level agreements (SLAs) that are applicable for
different stages of the buying journey, personalize communications, and measure the impact of marketing
programs.
Walking Your
Predictive Scoring Grid
Look for opportunities to use “OR” logic in your filters and triggers as you operationalize fit and
behavior scores. This will cover your bases and keep opportunities from slipping through the cracks.
http://www.infer.com/contact-us
A
1 2
4.2x 1.7x 0.9x
0.8x
0.7x
1.2x 0.4x
0.5x 0.0x
0.0x 0.0x
0.3x
15.2x
3.2x
2.1x
1.5x
3 4
B
C
D
Behavior Score
Likely to convert within the next 3 weeks
Fit
Score
Good fit to
buy your
product
http://go.infer.com/predictive-playbook/prioritization
6. About Infer
Infer delivers predictive business applications that help companies win more customers. It leverages proven
data science to rapidly model the untapped data sitting in enterprises, along with thousands of external
signals from the web. Customers include high growth companies like AdRoll, Cloudera, Concur, New Relic,
Nitro, Tableau, Xactly and Zendesk. Headquartered in Palo Alto, California, Infer is funded by leading
investors, including Redpoint Ventures, Andreessen Horowitz, Social+Capital Partnership, Sutter Hill
Ventures and Nexus Venture Partners.
Mountain View, CA
www.infer.com