Nominations for the 2024 GTM Awards, sponsored by TrustRadius and Pavilion, are open! Now through July 19, nominate top go-to-market leaders, teams, and companies in the B2B tech sector.  Learn more.


Predictive for ABM: Interview with J.J. Kardwell, President and Co-Founder of EverString

Megan Headley
Megan Headley
November 20, 2015

Predictive for ABM: Interview with J.J. Kardwell, President and Co-Founder of EverString

JJ Kardwell EverString

EverString is part of a relatively new space called Predictive Sales Analytics (aka Predictive Marketing, or predictive analytics for sales and marketing). We’ve found that many people are interested, but not yet ready to buy—honestly, some aren’t quite sure what predictive sales analytics actually is. We talked to J.J. Kardwell, President and Co-Founder of EverString, to better understand their positioning and the evolution of the space. EverString is unique in the space in that it’s architected for Account-based Marketing, designed around the account object rather than the lead object. Kardwell gave us an in-depth view on how buyers new to the space should be evaluating solutions, and how users already familiar with predictive lead scoring can be driving more value with other applications of predictive. 

Introduction to Predictive Analytics for Marketing & Sales

How would you characterize the space for readers who are early in the buying cycle, perhaps just starting out with their research?

A year and a half ago (when we started actively selling), category awareness was still very, very low. My sense is that nationally, awareness was virtually zero, and even in Silicon Valley it was less than 50%. Now, category awareness in Silicon Valley has grown to more than 80%, especially for tech and SaaS companies.

But for the rest of the country, I believe that category awareness is still below 50%, In Silicon Valley, adoption is approaching 10%; nationally adoption is still less than 1%. The typical person who will be using these technologies 3-5 years from now currently doesn’t yet know about the category and what can be accomplished with predictive marketing.

What kind of interest and inquiries do you get from buyers interested in the space?

The four questions I hear from buyers most often are:

  1. Can the product actually do what you say it can do? (However, this question is becoming much less common, as our product and the category are becoming more established.)
  2. What will this cost?
  3. What will the strain be on my organization, in terms of implementation and integration?
  4. How is EverString different from other predictive companies? (That is only coming from those people who are far enough along to have awareness of competitive offerings.)

History & Development of the Market

Can you give us a sense of the history of the space?

There are companies that self-identify as predictive marketing for marketing, and which have been around for almost a decade, but those are businesses which started as consulting companies. When pure SaaS companies like EverString entered the market a few years ago, these older predictive consulting firms tried to reposition as software companies. Most of the true SaaS companies in this market, which were born as pure SaaS and pure data science companies, started between 2010-2013. These are the companies that are out now out front in defining the market.

Foundation: The Potential of Applied Data Science

Before predictive marketing SaaS companies emerged, some consulting firms saw the potential of using applied data science to tap into and tie together internal data sources, giving marketers insights they wouldn’t otherwise have had. In a simple sense, for a marketer who is interested in doing segmentation, the application of data science enables segmentation on tens of thousands of variables instead of just a few. The earliest custom-built predictive deployments by consulting firms cost millions of dollars, and were only affordable for Fortune 500 companies. Human beings with PhDs built predictive models over the course of a month or more, and then the client could filter data on new and existing customers through that model. This can be immensely helpful, even if you are just using the historical data a company already has. For example, by looking at simple things like past buying behavior and usage data, the customer could recognize early warning signals for churn. As an example, if a customer’s logins and spend are going down by 25% every year, then there is likely a problem.

Key Technology Transformations and the Emergence of SaaS

Eventually, some key transformations at the technology level (independent of any individual company) made it possible for pure SaaS companies to emerge in this category.

One factor was the emergence of cost-effective on-demand cloud computing (from companies like SoftLayer, AWS, Rackspace and VMware vCloud Air). This meant startups could leverage that physical infrastructure to get started much more cost-effectively. In our business, if you’re doing predictive right, and you’re doing it at massive scale, the infrastructure strain becomes significant: crawling for data, high-speed compute, data storage, etc. During our seed round in 2012, we raised $1.7 million (we have since raised an additional $77 million). A decade ago, a new predictive company would have spent most of that money buying servers, without accomplishing much in terms of building the product or the business. In that regard, cloud hosting and infrastructure-as-a-service were foundational in the development of predictive marketing SaaS offerings.

The second foundational element is parallel computing, MapReduce and the emergence of Hadoop as a mature technology layer. That maturation of Hadoop happened in large part through the emergence of businesses like Hortonworks and Cloudera, which made Hadoop technologies more usable in an enterprise-grade setting.

The third major contributor to the emergence of this category is innovation in data science methods, and particularly in the areas where EverString has been leading the industry in innovation around automated feature selection and automated model building. What currently takes a consulting company one to three months to do, can now be done in a matter of hours by a fully automated platform.  Turning data science into software is one of the most difficult things to do in our market.

With these developments, a market that had initially been consulting-driven is giving way to true SaaS.  I have seen this trend replicated across many industries – early market entrants with consulting or “technology-enabled services” business models almost invariably get displaced by pure software solutions.

From Predictive Lead Scoring to Full-Lifecycle Predictive Marketing

What emerged from the early days of predictive was a category that, a year and a half ago, still self-identified as “predictive lead scoring.” EverString really started our business 15-18 months ago; before that, we were in R&D and platform development mode. Until February 2014, our entire team was just 12 data scientists and developers.

We’ve set out to change the way people think about this space.  “Predictive lead scoring” is a narrow and terrible moniker for this category. Although some companies still use that term, our view is: If you believe in predictive—if you think applied data science works—why would you only use it inside your existing pipeline? Why would you limit the application of predictive to only within your CRM or marketing automation system, solely to reprioritize existing leads? Why not apply it outside of your pipeline as well, to find completely new prospects?

The second problem with the notion of “predictive lead scoring”, is that you should not just focus on the “lead” (i.e., contact or person).  The best-in-class predictive deployments start by evaluating companies (i.e., accounts).  Focusing on the company before the person enables you to extend predictive outside of your existing pipeline, to find new companies that look and act like your best customers.

To put this in perspective, in the U.S. there are approximately six million companies that are large enough to have a website. Typically, the customers we talk with have approximately 50,000 to 100,000 prospects in their CRM and/or marketing automation. What about the 5.95 million companies that are not in your database? Why not use applied data science to have a view on them as well? This type of proactive “predictive demand generation” is more difficult to do than simply reprioritizing your existing pipeline.

Simply repackaging existing data about your contacts and their transaction histories from your marketing automation and CRM systems is not enough anymore. We collect massive amounts of external data, which enables our customers to have a point of view on every single potential prospect in their addressable market, regardless of whether that prospect is already in their pipeline. You shouldn’t be marketing to every company, but you should have a point of view on everyone—then you can make an informed decision about which companies and people are worth marketing to, advertising to, and selling to.

The Predictive Space Today: SaaS vs. Consulting

Taking a broad look at this market and its evolution, I see a distinction between SaaS and consulting businesses.  It’s very easy to tell them apart. Don’t ask the vendors, ask their customers: “How long did it take to build and deliver the predictive model?”  If the answer is 30 days, or 6-8 weeks, it is painfully obvious that humans are building the model manually. That is not a software company. Regardless of positioning or glossy messaging on websites, the time to build a model and fully implement is one of the best ways to cut through the spin and test if a company is actually trying to sell you software or consulting.

I also see a distinction between companies that take a comprehensive view of predictive’s potential, versus those that only deliver lead scoring, whether because of technology limitations or a narrow belief. There is a very small universe of people doing anything else beyond lead scoring with predictive.

Beyond Predictive Lead Scoring

From your perspective, what can/should predictive analytics do, beyond lead scoring?

The main uses of predictive for sales and marketing fall into three buckets:

1. Predictive scoring (versus only predictive lead scoring)—if you want to be an account-based marketer, you should be scoring accounts (i.e., companies), not leads (i.e., people). You should be looking for new accounts that look like accounts that have converted. Therefore, most of our customers are actually doing account scoring. After doing account scoring, you can also do person-level scoring (i.e., lead scoring) based on title and engagement.

Our platform is the only one on the market built with an account-based architecture. Others were built with a lead-based structure, essentially as extensions of marketing automation systems, and are therefore trapped in the lead object. A narrow focus on the individual person can be a dead end. What if you have a lead that is a person with the right job title, but the company 100% isn’t a good fit to become a customer?  That is a dead end, no matter how “good” the person’s job title is.  Conversely, what if you are talking to the right company, but the wrong person? That isn’t a dead end, as you can always append additional contact information for the right people if you are starting with the right account. If you aren’t focused on the right accounts, nurturing and retargeting are a waste of time, and no amount of selling matters. This is why starting with account-based scoring is so important for B2B marketers, and why it is so alarming that most predictive companies are pushing “lead scoring” on every single customer.

In terms of predictive scoring, it’s very clear how we differentiate. We are the only predictive marketing company with an account-based architecture.  While account-based marketing (ABM) has been around for years, the broader market wasn’t focused on it until the past 12 months.  Marketers started realizing that ABM is a very real thing when they saw Jon Miller leave Marketo to start Engagio, and when people began hearing marketing thought leaders like Maria Pergolino at Apttus and Craig Rosenberg at TOPO publicly explaining ABM.    Three years ago when we starting building the EverString platform, we weren’t using “ABM” terminology, but we did very intentionally set out to provide customers with an account-based platform for predictive marketing.

2. Pipeline building— our product for building new pipeline is called Predictive Demand Generation. This involves using predictive analytics and data gathered from the web to find completely new prospects that are not yet in your pipeline. Within predictive demand gen there are two subcategories: intent-based predictive demand gen, and fit-based predictive demand gen.  Some companies only offer one or the other approach, but we offer both, as the optimal way to evaluate prospects is using the three dimensions of fit, engagement and intent signals at the same time.

Intent-based predictive demand gen looks at person-level behavioral data. Intent is essentially engagement which is not with you.  In other words, it involves people visiting your competitors’ websites, reading blogs related to your market, and searching for relevant terms. Nobody has access to more intent data than we do, as we needed to build this depth to support our new Predictive Ad Targeting product, and the benefits of that data flow benefit all of our customers. Engagement can be long-tailed, whereas intent (companies in market, ready to buy right now) has a much shorter fuse. Intent signals have very limited reach, since there is obviously a very small percentage of prospects who are truly in market trying to buy at any time, but those signals can be very powerful.

Moving across the spectrum, we look at each of intent data, engagement data, and fit data. Fit-based data provides the broadest coverage of the three, so is very impactful from a reach standpoint.  Fit data includes traditional firmographic information (e.g., company size, industry and geography), tech stack data (i.e., what products and technologies are they using), and person-level data about title, background skills, etc.  Your sales team probably shouldn’t be on the phone with anyone unless the predictive platform has automatically qualified the prospect account for all of those key things. Otherwise, it’s not a good use of human time—it’s only a good use of machine time.

3. Engagement—our latest product, which drives new engagement is called Predictive Ad Targeting. This enables B2B marketers to identify ideal prospects (both companies and people) using our predictive platform, and then proactively deliver ads to them.  This gives B2B marketers unprecedented control over the top of their funnel.

EverString Positioning & Differentiators

So how do you differentiate from competitors?

1. EverString has the only platform with an account-based architecture. Most predictive products are built as an extension of marketing automation, and can only score individual people (i.e., leads) rather than both people and companies (i.e., accounts).

2. We are the only full funnel, full lifecycle platform.  This means that we can apply predictive not just for prioritizing existing prospects inside your pipeline (i.e., scoring), but also to optimize and control the upper funnel by building completely new high-propensity pipeline.  I encourage potential users of predictive to ask about more than just scoring. Our viewpoint is that the best use of predictive depends on your specific needs. Scoring is great if you are a B2B company that sells like a B2C company, with millions of leads pouring in, and a high volume of free users who you need to prioritize to identify potential for upsell to a paid offering. In that case, you really need prioritization more than anything; that is the single highest impact thing you can do with predictive.

But that is not the profile of most B2B companies. We believe that approximately 98% of B2B companies suffer from having too few prospects.  For them, even if they have good inbound lead flow, they will hire as many sales development reps as needed to call every single new lead. If you’re already calling every lead, scoring is less impactful. For 98% of companies, the biggest impact they can have with predictive is actually pipeline building through predictive demand generation or predictive ad targeting. We increase conversion rates with our predictive scoring product, and we increase pipeline with our predictive demand generation and ad targeting products. Every marketer and every salesperson wants a higher conversion rate and a better pipeline filled with more prospects that look like their best customers—that’s what we offer.

3. Our platform is fully transparent. Most platforms on the market are “black box” models. Most thinking marketers want to look into their model and see how individual signals impact performance. They want to understand why the model is working, not just blindly trust it. They want to be able to see that the model has figured out on its own what matters to them.  If the model automatically figures out things that the marketer knows are true, that builds trust in the model. The transparency might also introduce new dimensions that cause marketers to say “I never thought about that—we do well with companies that have a massive Twitter following, are using AdRoll for retargeting, Marketo for marketing automation, AppDynamics for APM, and that have a C-level management person who used to work at Google.” We help marketers uncover the confluence of factors — the compound variables — that are impossible to identify with just intuition.

4. EverString is a true SaaS offering. Ours is a market where a lot of people assume that every vendor that says it’s a SaaS business must actually be one. The reality is that not all of the predictive companies are truly selling software. Predictive marketing is not a workflow business like marketing automation or CRM.  Work flow tools make it easier for users to complete a given set of tasks with the fewest clicks, and in the least amount of time. That’s helpful, but the shortcoming is that you could put a million leads into those systems and not close a single deal if the leads are the wrong targets. EverString, on the other hand, provides an intelligence layer that helps make sure you are identifying and focusing on the right prospects. EverString runs like an infrastructure layer — you set it up, and then it runs on an automated basis.

In the workflow world, buyers evaluate products based primarily on user interface. For predictive marketing, people should be evaluating products just as much based on the application layer – not just the presentation layer. We believe that our product has the best presentation layer on the market, but we think it’s important for buyers to drill down, and ask how much is being done by software on the back end. Are your models being built automatically by software, or are you waiting for humans to manually build a model? In our case, we can build a model in two hours using our automated platform, and we can have customers live inside of two days.  This contrasts dramatically to the non-software predictive vendors who take a month or longer to build a single model, as they are doing it manually instead of using software.

Are there other indicators, besides time to build for the models?

Aside from model build time, another way to tell whether a company is delivering software or consulting is what percentage of people in their company are professional services? In predictive, if more than approximately 10-15% of the employees are providing professional services, it’s likely not a SaaS business.

All of this matters for customers. While every single model EverString builds is 100% customized for each customer, there are no non-recurring engineering or custom build fees — it’s a push-button integration that can be completed in five minutes, and then the model is automatically built by the platform over the next two hours. With EverString, implementation is a single 30-60 minute call to talk about what the customer wants to predict (e.g., raw lead to marketing qualified lead, or marketing qualified lead to opportunity, etc.). Once people understand the difference between a predictive software solution and a consulting engagement, they have a whole new perspective on our value proposition and on the market.

How do you see market perspectives changing?

We’re still at a stage in the market where most people don’t know what predictive marketing can do for them. Unfortunately, if they’ve heard of “predictive,” they’ve likely only heard of “predictive lead scoring”, and they are building their evaluation framework around that. We want to educate the market about the full array of options for leveraging predictive for sales and marketing.

We have grown from 12 employees in March 2014 to 126 employees today. During the past 15 months, we’ve expanded from a fraction of the size of our competitors to double the size of most. People ask all the time how have we have accomplished this? Literally by using our own EverString Predictive Demand Generation product to identify new high-propensity prospects and to feed our growing sales team.  EverString’s own growth is the most obvious proof for why marketers should want to do more with predictive than just lead scoring.

EverString Customer Base

Who is your buyer? Who at the company is making the decision to use predictive?

70% of the time it’s marketing: the Director of Demand Generation, the VP of Marketing, or the CMO. 10% of the time it’s Operations: either Marketing Ops or Sales Ops. 15% of the time it is Sales. 5% of the time it’s advanced analytics – internal data science teams.

What does adoption look like, among the three use cases/products?

We just launched Predictive Ad Targeting a week or so ago, and we have three customers using it already. Approximately 80% of our customers are using both Predictive Demand Generation and Predictive Scoring. The remainder is split pretty evenly between those two use cases. We pioneered the capability to use predictive analytics for pipeline building with predictive demand generation.

The reality is that once customers are succeeding with one use case, many see the power of predictive and want to use everything that we offer.

Who’s the end-user of EverString? Who helps with the integration?

It’s typically the Demand Gen marketer or Marketing Ops. EverString really is a product that runs itself once it is integrated. The integrations takes 5 minutes. We then spend an hour on the phone with the CRM and/or marketing automation admin, and then the platform does the rest.

Can you give us a view on your customer base?

We are quickly approaching 100 customers. We sell to large enterprises and mid-market companies. Our customers include Comcast Business, IBM, Microsoft, Apttus, Zenefits, 8×8, Hortonworks, and many others. Our customers are all B2B companies, but not exclusively in the tech industry. Our view is that predictive should work not only throughout the funnel, but also in every industry. We have many customers in non-tech industries, like insurance, business services, telecom, etc. EverString’s platform works equally well in any industry, and with B2B companies of almost any size.

Predictions for the Predictive Space

Do you see the predictive functionality being used in other ways in the future?

Yes. The models we build are integrated at the CRM, marketing automation, and website levels (we have our own pixel running on our customers’ websites). If you zoom way out, this is about determining the profile of the ideal customer (contact, lead, or account), based on past conversion, and then selecting the right audience both inside and outside of the current pipeline. Predictive provides a way to drive and monitor conversion throughout the funnel, which enables full-lifecycle, closed-loop attribution and reporting. You will also see the power of predictive more broadly in organizations, across business functions outside of sales and marketing, as well as also closer to the edge, at the desk of individual salespeople.

EverString has had the most ambitious and productive R&D capability in this sector. We were the first to launch both Predictive Demand Generation and Predictive Ad Targeting. You’ll see some other big innovations from us in the next few months.

“Prescriptive analytics” is a buzzword that comes up a lot in this category. What are your thoughts on that term—does it apply to EverString?

Predictive entails predicting what will happen, whereas prescriptive involves suggesting what you should do. The market is on buzzword overload. People are trying to wrap their heads around predictive, and using the term “prescriptive” just adds excess complexity in most cases. That said, we’ve already delivered the majority of the capabilities needed to make this a reality.

If you’re using predictive right, you should absolutely be doing it in a prescriptive way—not just delivering the calculation that something has 3x the chance of conversion. The notion of pointing to the “next best action” isn’t a major evolution – it’s a minor incremental development at the edge of current predictive deployments.  The true next major evolution of this category, beyond predictive and prescriptive, will be to full artificial intelligence enabled decision making and execution.

Prescriptive involves recommending an action. For example, “Based on history and trends, you should send this prospect this piece of content or call them right now, or engage in these next three steps.” The next level beyond that is full artificial intelligence enabled decision making. For a given budget, this would allow users to push a button, identify the ideal audiences and also optimize execution across multi-channel delivery. The predictive model will learn, optimize, and execute all campaign spend. But that kind of pure AI decision making and execution is still a few years out.

One other issue with prescriptive is that deployments can very quickly get bogged down at the edge where recommendations are being made.  The problems arise where software crosses over into human hands-on areas, like content building, creation, and curation. In the world of “next best action” there is a certain distance you can go purely automated, but in order to fulfill the true dream of prescriptive you need a human element—for example, to write and tag content in a way that the software system can intelligently grab and distribute it to micro-segmented audiences.

It’s a major investment of manual time and energy to truly get full value out of the prescriptive software.  The vast majority of B2B companies we talk with that are considering predictive marketing are primarily focused on expanding their pipeline with Predictive Demand Generation or Predictive Ad Targeting.

Anything else you’d like to add?

Three things:

1) People often ask how predictive is different from business intelligence (BI).  BI is 100% backward-looking; it’s all about observing what has already happened. Predictive is the first application of analytics to actually be forward-looking and provide a point of view on the future.

2) In a market this early, customer success matters a lot. The challenge is that most predictive companies have a small-company mindset; they’re focused almost exclusively on winning new customers. Unlike a product such as CRM – which most customers have previously deployed, and have significant familiarity with – almost no one has deployed predictive marketing before, and they have no idea what to expect. Customer success is therefore crucial to this category, but we see very few vendors approaching it this way.

In order to facilitate customer success, it’s important to be very clear up front about the customer’s pain points, needs, and expectations. Why are they using predictive, and what do they hope to get out of it? It’s also very important to determine baseline performance. Everyone needs to know what the conversion rate was before we deployed predictive, so that we can quantify improvement. Relative to non-predictive sources of prospects, we’ve seen situations where our predictive prospects have 30x better conversion performance. At the very least, customers are seeing 3x or 5x improvements.

3) Lastly, it is important for companies considering predictive to understand how quickly it can be deployed, how quickly they can experience improved performance, and how broadly they can drive impact.  While some of the predictive vendors with a consulting engagement model can take a month or two to build a model, customers using EverString’s true SaaS platform can have a model built in a couple hours, and can be live on the platform in under two days.

We have built a platform that makes it very easy for customers to try predictive in a rapid and low-cost way.  Perhaps most importantly, customers who are making the investment of time and money in deploying predictive should insist on being able to use it not just to prioritize their existing leads with scoring, but also to find and engage with new prospects from outside their funnel by using products like Demand Generation and Predictive Ad Targeting.

About the Author

Megan Headley
Megan Headley
Megan leads Research at TrustRadius, whose mission is to ensure TrustRadius delivers high quality, useful and, above all, trustworthy user feedback to help prospective software buyers make more informed decisions. Before joining TrustRadius, Megan was Director of Sales and Marketing at Stratfor, where she was in charge of growing the company’s B2C revenue stream through email marketing and other channels. She enjoys traveling, reading, and hiking.

Sign up to receive more buyer resources and tips.