Evolution of Predictive: Interview with Jamie Grenney, VP of Marketing at Infer

Infer’s flagship product is a predictive analytics platform for marketing and sales, which builds predictive models based on a company’s historical pipeline and current customer data. The predictive sales analytics space is heating up, and as vendors become more competitive, they’re developing differentiators. We talked with Jamie Grenney, Vice President of Marketing at Infer, about their new “stealth” product, Prospect Management, which can work with or without a predictive model. Infer Prospect Management is a lightweight app that works with Salesforce, marketing automation, and/or data from a predictive model to produce recommendations about which prospects to target and which content to use. Prospect management helps ops teams, product marketers, demand gen teams, and sales reps get a more intelligent view of segments, personas, and profiles. Grenney also discussed the evolution of predictive technology and weighed in on the possibility of “self-driving cars” (AI-powered marketing and sales ops) in predictive’s future. 

Introduction to Infer

What’s new at Infer?

Over the last year, $242 million in venture funding has flowed into B2B predictive marketing. That is great validation that we’re on the edge of a mega-trend that could be every bit as big as cloud, social, or mobile.

At Infer we feel fortunate to have gotten out front and had time to build up tribal wisdom. By working with lots of forward-thinking companies, we’ve been able to learn the edge cases, improve the technology, and develop playbooks for success. Just as we saw with Salesforce, the community you attract and the success of your customers is the catalyst for growth.

About a year back we had a realization that scoring alone was not enough to win this market. From working with customers we knew that you can have a great predictive model, and that is critically important, but a score by itself is not enough. It is the applications of the scores that unlocks value.

So to get ahead of the competition, we made a strategic bet. We split off a big chunk of our engineering team to build a new stealth product. One that takes predictive models to the next level, but interestingly enough, doesn’t require predictive models. It accelerates a customer’s time-to-value, elevates the plateau, and is applicable to companies of all sizes. It’s a lightweight platform that runs alongside Salesforce or your marketing automation app, pumping out intelligent recommendations on how to hit your number and what messages to use.

Can you give us some brief background on Infer, for readers who aren’t yet familiar with the company?

Our core predictive lead scoring product helps B2B companies analyze buying signals and predict which prospects will go on to become great customers. To do this we look at two dimensions. Who is a good fit to buy your product; do they look like your existing customers? And then we look at who is showing buying intent and likely to buy soon. We figure that out by analyzing data collected in your marketing automation application and Google Analytics.

Having an accurate score is the first step, but as I mentioned earlier, it is the application of the score that unlocks value. Across our customers we really see five core use cases for predictive scoring: filtering, prioritization, net-new leads, campaigns, and nurture.

• Filtering and prioritization are important for companies that have too many leads. For example, Optimizely has lots of trial leads and HubSpot gets tons of content marketing leads.

• There are other companies where inbound is important, but they need to complement it with an outbound strategy. For example, companies like Xactly or Belly. They use predictive to prioritize the right accounts to focus on and get more efficient at finding net-new contacts.

• Finally, predictive helps marketers. Campaign teams are using a new metric called “Cost Per Good Lead” to test & invest faster, and they’re using predictive to surface gold out of nurture when behavior scores spike.

Customer Base & Common Use Cases

Are SaaS customers your sweet spot, in terms of your customer base?

Similar to Salesforce and Marketo, we decided to start by focusing on high-growth, mid-market tech companies. Our earliest customers included Box, Tableau, and Zendesk. Over the years, we’ve added many other great logos and we have started to expand into new verticals.

One thing that differentiates Infer from some of the other vendors is our DNA, or what we’ve learned from our prior experiences working at Google and Salesforce. From the very beginning, we have focused exclusively on being a products company. We wanted to figure out the use cases where predictive could unlock the most value, prove that we could deliver repeatable customer success, and scale to meet the needs of thousands of companies. We’ve avoided solutions that rely heavily on consulting services, because we believe that one off solutions prevent you from achieving network effects and delivering value quickly.

Using this approach we’ve been able to grow with all kinds of companies — from those with less than a hundred employees, through to large successful public companies.

Can Infer be used beyond the use case of lead scoring? Are there other areas where predictive capabilities can be leveraged?

Predictive Lead Scoring is where we got our start because that’s often the easiest spot to unlock value. When someone first fills out a form on the website you know very little about them, yet Infer can bring thousands of external signals to the table. With nothing more than an email address, we can accurately predict which leads should be rushed to sales and which should be filtered out.

Over the years we’ve added other types of models. We offer fit models and behavioral models, we can score different objects like accounts and contacts, and we can help companies apply different lenses so they can use predictive for various geographies, market segments, or verticals.

The other thing I should highlight is that Infer Prospect Management is a huge leap forward. Most predictive vendors give you 4 segments: A-Leads, B-Leads, C-Leads, and D-Leads. But with our new product, we’re letting people define their own segments, personas, and states based on thousands of signals that we collect and analyze.

How do you view the predictive space, in terms of evolution and maturity?

I can imagine that for a buyer it must be overwhelming to have dozens of vendors saying they all do something similar. There is a lot of hype in the market.

To help navigate the space and see where the technology is headed, we’ve drawn an analogy with the evolution of maps. If you think back, we went from AAA maps, to interactive maps on our phones, to GPS apps with turn-by-turn directions, to real-time intelligent recommendations in Waze. Now off in the horizon, we have self-driving cars. B2B companies are going through a similar progression. An accurate score is the first step, then it is about the applications of the score, and the next breakthrough is intelligence. Is the system able to proactively recommend ways to unlock value? This analogy is helpful to reference, because many vendors are still at level one, talking about the signals that go into their score. There are a few who can go deep with use cases and provide strong customer success stories. But really, you’d like to be at level three. You want a platform that can provide intelligent recommendations you can act on with a couple clicks, and unlock measurable ROI.

New Product Launch: Spotlight on Prospect Management

How does your new product change the game?

To start, Infer Prospect Management has connectors that let you gather all your data in one place. Across Salesforce, Marketing Automation, Web Analytics, and External Data, that alone is no small task.

With this data at your fingertips, you can create a set of profiles and workflow rules that let you synchronize your strategy across automation systems. All of your customer segments, personas, and states become data-driven. Instantly you can tell where you’re over-invested or under-invested, and you can see what messaging or action is most effective at moving prospects forward.

Prospect Management is like the brains of your operation. When you’re using AI to assist employees and improve the customer experience, everything runs smoother. People get the right message at the right time. And you’ve always got a steady stream of actionable data-driven recommendations to help grow your business.

What departments can use Prospect Management, and how are their use cases different?

This is an application that every customer-facing employee in the company can benefit from, but especially those in sales and marketing. For example;

• Someone on your operations team, either marketing or sales, might want to understand which segments are over and under-invested.

• A product marketer might want to build a persona to measure its purchasing power and analyze the typical customer journey.

• You may have folks who are focused on demand generation and campaigns. They finally they have the ability to define narrow segments, assign specific actions, and measure what is most effective at moving buyers forward from one play to the next. Of course, once you’ve found that perfect profile, Infer makes it easy to go out and get more of those buyers to fill your funnel.

• If you’re a Sales Rep or an Account Executive, you’re going to be the beneficiary of the recommendations. Infer will help you prioritize your work to reach your number faster. Or it will recommend the right piece of advice, the right action, at the right time, so you can address the customer’s need and move them forward in their buying journey.

What does the implementation process look like?

Most companies will start with Prospect Management because it’s super simple to set up. All you need to do is authenticate into Salesforce and you’re off and running. When you login, all your leads, contacts, and accounts will be matched to Infer’s library of external signals. You’ll also be able to add other connectors such as Google Analytics or Marketo.

While many customers will be able to get value out of Prospect Management as a stand-alone product, you can also add a custom Infer predictive model. As long as you’ve got sufficient data (a couple hundred examples of good and bad prospects), it’s pretty easy for us to build a model. In addition to Salesforce access, we just need to do a quick 30-minute phone call to understand your processes and make sure we’re aligned on what you’re trying to predict. From there we’ll use machine learning to build a set of models, and zero in on which one is most effective. About a week later we’ll meet to present the model, demonstrate its accuracy, explain the signals it uses, and recommend ways to apply the score (the one or two playbooks you should start with). Our goal with every customer is to begin with a vision of what success looks like, and within 60 days prove out that success story.

Prospect management sounds similar to lead management, which is something we hear about from marketing automation vendors. How does Prospect Management compare to marketing automation?

We view them as very complementary. Marketing Automation Platforms play an important role in gathering data and executing workflows. We have no plans to focus on applications like landing pages or robust mass email tools. Our platform was designed as an open platform, designed to be lightweight, and designed to share intelligence (e.g. profile tags, messages, data) across all the tools in your toolbox.

Infer’s Predictive Platform

Do you work with the same buyer and admin on the predictive side?

Yes, oftentimes the marketing or sales ops person is the person we work with day-to-day. They are often the people tasked with overseeing scoring and predictive analytics. There can be lots of people who champion an application like this, though. It is an important strategic initiative for VPs and CXO execs.

How many customers do you have on the predictive side?

We have deployed many, many customers on our predictive models so we have a good bit of experience working with companies. And as I mentioned earlier, a good number of them are companies like Box, Zendesk, and Tableau, who started working with us at a young age and have grown up to become public companies. We think that the number of customers we’ve worked with and the caliber of companies is an important competitive differentiator.

What does onboarding look like at this point?

Predictive models are relatively simple from a setup perspective, especially compared with other marketing technology companies might have implemented. For example, if you’ve rolled out a marketing automation platform you know you need to migrate your data, get code on the website, change the culture of the campaigns team, code up a bunch of emails, etc. before you can generate measurable value.

With a predictive model, the steps are simple. We need access to your Salesforce data, so if you have the keys to do that, authenticating in is very quick. We’ll have a quick conversation (30 mins.) to understand business processes and goals, because we need to know what you’re trying to predict so that we can train the model in the right way. Then we build the model, which is fast and automated since it’s all done using machine learning on our end. We next have a data analyst check the models because we believe it’s important to have that human touch point.

When we review the results with a customer, we take them through historical back testing, which shows them how accurate the model was at predicting winners. We’ll take them through the signals within the model, as well as a strategy for rolling it out, so that they can understand what their success story looks like, including the metrics we’re trying to move in the first sixty days. For example, if we can see that there is too much energy going into C and D-Leads, we advise them to shift that energy into higher value activities. Because it takes a data-driven approach, we can project the impact on conversion rate, win rates, or average deal size.

Looking Forward to the Future of Predictive

What advice can you give buyers who are unfamiliar with predictive?

With any technology, the simplest piece of advice is to look for the success stories. Does the vendor you’re talking to have a reference customer that you aspire to be like? Do they have a detailed story around the playbook they implemented? Do you have confidence that they’re going to be a trusted advisor, routing you around pitfalls and towards proven ROI?

Especially with a new space where there is lots of hype, the surest path to success is to follow in someone else’s footsteps.

Going back to your GPS analogy, is there a self-driving car in your future?

Just like self-driving cars, the move towards autonomous marketing is inevitable. It may be five or ten years off, but we’re already seeing visible examples of it in the advertising space with programmatic bidding. Clearly, there are tasks where humans are better equipped to make decisions, but for many repetitive tasks, such as researching a lead, the AI approach is infinitely more scalable and far more accurate.

The opportunity for forward-thinking marketers is to figure out how to leverage AI in their careers. If you can become the expert, defining the profiles, the actions, and the next best actions, you’ve got an opportunity to make a tremendous impact. That’s really our goal. We want to find great champions and make them heroes within their organizations.

To share your own user insights about Infer, write a review on TrustRadius.