ZenIQ Account Based Marketing System Maps Buying Centers, Finds Data and Exection Gaps, and Recommends Actions to Fill Them


Share on LinkedIn

When I first starting thinking about Account Based Marketing, I assumed that an ABM system would let marketers replicate at scale how sales teams manage key accounts: that is, to analyze each account in depth, set goals specific to that account, and then execute against those goals. But most vendors serving the ABM space have taken a much narrower approach, either in providing data about accounts, managing campaigns against externally-built account lists, or providing account-level metrics such as coverage, engagement, and funnel velocity. Vendors who offered these things told me that the account-specific planning I imagined wasn’t practical and, in fact, was rarely done even by account teams in sales.

I was disappointed but figured it was just another case of expectations outpacing reality.

Then I saw ZenIQ.

ZenIQ assembles account data from a company’s CRM, marketing automation, and Web systems; supplements this with account and contact information from external sources; assesses the current state of each account; and takes actions to improve that state. At present, the actions are chosen by rules set up manually by marketers – although even this is a step ahead of having marketers directly assign accounts to specific campaigns.  Later this year, ZenIQ plans to release machine learning-based recommendations that will, in effect, generate the rules themselves.  Even automated recommendations in place, ZenIQ won’t select your target accounts or execute the recommended actions. But tools for both of those tasks are widely available and, when it comes to execution, most companies don’t really want to replace their existing email, Web, CRM, and other execution systems. So ZenIQ comes about as close as anyone could want to to providing a complete ABM system.

Let’s take a closer look at how all this works:

ZenIQ starts by importing accounts and contacts from a company’s marketing automation and CRM systems, including static attributes and behaviors. It also places a tag on the company Web page to capture visitor behavior directly. The system applies sophisticated matching to unify contact data and to link contacts to accounts. It then enhances the contact and account data with attributes, events, and intent from the usual ABM data vendors. 

Now things start get interesting.  Contacts in each account are assigned to one or more “buying centers” and then classified by their role and importance within each center. This classification relies on machine learning to map titles and interests to standard buying roles such as influencer and decision-maker. ZenIQ next examines each buying center to find coverage gaps – that is, standard roles for which no contact has been identified. The system then fills those gaps with contact records from external sources. This is the sort of work you’d previously have needed a pretty smart sales rep to handle properly.

Once the machine learning pieces are fully operational, ZenIQ will look for accounts with unusually low message volume and engagement, relative to all accounts for that client.  It will also infer contacts’ personal interests, channel preferences, and optimal message frequency from their behaviors in marketing automation, CRM, and the Web site. Automated classifiers will tag CRM, Web, and marketing automation activities across multiple dimensions (channel, engagement level, initial vs later contact, etc.), assign stages to opportunities (early, middle, and late) and find correlations between activities, gaps, and outcomes. These correlations will be the basis for recommending the next best action for each account.

This sequence sounds almost entirely automated, but ZenIQ recognizes that human input is needed (at least for now) to keep the machines from making foolish mistakes.  So classifiers will undergo a training period during which marketers review and correct their results.  Similarly, marketers will review the system’s recommended actions and approve them before they are passed on for execution.  

Marketers will also design the actions themselves.  These will be processes that execute in ZenIQ and, mostly, in external systems.  For example, a typical action might be to buy new contact name and add it to a marketing automation database. Users can create such actions in ZenIQ today and embed them within “recipes” that also contain a rule for when they execute.  Recipes can be driven by real time events (“send a task to CRM if a decision-maker requests a meeting”) or by scheduled processes (“check daily for new accounts without a decision-maker and fill any gaps you find.”)  They can add contacts to campaigns in Salesforce, Marketo, Eloqua, Hubspot, Pardot, or other systems with a standard API connection. Those systems could in turn feed other channels such as display advertising. ZenIQ will also report on coverage, reach, engagement, pipeline movement, and results over time.

During normal operation, ZenIQ will receive regular updates from CRM, marketing automation and Web tags, and react appropriately. Think of it as a smart little robot supervising every account relationship and suggesting the right thing to do in each situation. In short, it pretty much matches my original expectation of what ABM would be.  Score one for reality (and ZenIQ’s creators).

ZenIQ was founded in 2015 and released its product in early 2016. It is currently out of beta but not quite officially launched. Pricing starts at $36,000 per year for 6,000 accounts, plus $50 per user per month. There are additional fees for more accounts and for downloaded contact names. The company reports 11 current clients.

Republished with author's permission from original post.


Please use comments to add value to the discussion. Maximum one link to an educational blog post or article. We will NOT PUBLISH brief comments like "good post," comments that mainly promote links, or comments with links to companies, products, or services.

Please enter your comment!
Please enter your name here