Advancing Human-Computer Communication to Maximize Sales

Pirtle, Bryan1-VC

Bryan Pirtle, SDM ’13

By Bryan Pirtle, SDM ’13

The challenge: In today’s fast-moving, dynamic business environment, it is tougher than ever for businesses to reach the right clients in a meaningful way and achieve results. Sales and marketing professionals need highly specialized and effective technology to source clients, engage them, and close deals.

This is especially true for top-of-funnel outreach. Even if a sales/marketing professional has carefully selected targets, “cold call” emails are almost always deleted immediately. How can prospective customers be incentivized to engage with the material for mutual benefit?

The approach: Nova Labs was formed to address this challenge using an integrative, systems-based approach—coupled with deep knowledge of typical top-of-funnel challenges—to develop hyper-personalized email outreach technology that surpasses what is currently available to sales and marketing professionals.

First we conducted a massive information-gathering mission by interviewing dozens of lead users in companies across the globe. The goal was to understand how potential users in the sales/marketing field spend each day on the job. Findings produced a clear view of common and best practices and typical products already in use.

Next, we researched the market landscape to gain insight into what gaps exist at the leading edge of sales automation technology.

Finally, we created a software-as-a-service, business-to-business product designed to meet the new-customer acquisition needs of medium and large companies. The goal of this initial product, Nova, is to optimize email outreach for clients. Nova’s underpinning is a scalable personalization technology that can be extended into other domains and ultimately become a platform for future products. This technology will enable our customers to personalize not just email, but ads, engagement/re-engagement communications, upsell opportunities, and more.

The tools: The systems thinking mindset, technology strategy, and user-centered design—as well as system architecture and dynamics methodologies—that form the foundation of MIT’s System Design & Management program have provided many of the fundamental concepts on which Nova Labs built its core technology.

The product design/synthesis/feedback process included:

  • Transforming all of our documented experiences plus the synthesized information collected from the sales/marketing professional interviews into baseline product requirements;
  • Choosing and configuring the technology stack to be used;
  • Determining initial/future sources of information from the public Internet with which to construct meta-profiles for each prospect;
  • Creating a baseline scope for the personalization technology layer;
  • Designing the look and feel of user interfaces;
  • Designing the “glue layer” of how data flows to/from each interface, as well as usability/user experience;
  • Implementing all of the above in the chosen technology stack;
  • Sourcing beta clients and releasing the product to them; and
  • Feeding back client suggestions, concerns, and usability issues into the process to continue iterative design.

After analyzing and discussing the results of our data collection efforts at length, we were able to characterize our target market as follows:

  • The client: Sales development representatives, account executives, and personnel in inside sales roles.
  • What the client does: Spends five or more hours each day researching individuals on prospect lists to determine the right personal touch to add to an initial message to engage each person.
  • What the client currently uses: Static, template-based, single-email, and mail-merge (multiple-email, campaign-based) platforms that convert comma-separated values documents into bland, spammy-looking emails.
  • What the client wants: Dynamic, data-driven technology that allows for utilization of publicly available data and performance analytics to construct appropriate targeting language automatically and insert it into emails to capture each recipient’s attention and/or establish a personal connection—at scale.

We thus determined that a technology that combines the successful elements of existing products on the market with an automated personalization text-generation layer would have the potential to revolutionize sales and marketing top-of-funnel and save hours per day per professional.

To deliver this product, we created a robust and sophisticated software stack capable of asynchronously compiling and delivering the data needed for the personalization process and desired user experience. Our current stack with high-level relationships between components is shown in Figure 1.

NovaBlockDiagram

Figure 1. Nova Labs’ technology stack.

We decided to use contemporary open-source software built upon the Heroku platform as a service in order to achieve the following goals with our stack:

  1. Stability, gained by using robust, mature frameworks and software platforms;
  2. Flexibility, attained by picking and choosing the correct open-source tools and libraries for each product need; and
  3. Focus on application design rather than infrastructure design.

We built the primary user interface as a Google Chrome extension application. This design choice was primarily made to make use of the Google Apps mail client that most lead users preferred. We chose AngularJS as the front-end Javascript framework and “glue layer” to provide a more robust and streamlined real-time “desktop experience” in the web environment. The asynchronous and real-time data needs of the application were also simplified by the use of AngularJS. Figure 2 shows the real-time analytics in action in a campaign.

Pirtle_dashboard

Figure 2. Screenshot of Nova’s “campaign view” with analytics data.

Once a client uploads a list of target email addresses, the personalization engine is employed on the back end to search target sources on the Internet. The engine then constructs a meta-profile for each person using his/her email as the key. We chose several initial data sources, including social media data aggregator APIs as well as our own data scrapers and search technology. Proprietary algorithms are then used to construct the meta-profile and custom attributes—such as seniority, influence, playfulness, and international orientation—from the collected raw data.

Once the meta-profiles are constructed, the personalization engine uses natural language processing and machine learning as well as statistical analysis of past analytics data to determine the proper personalization snippets, including tone, to add to the email for each recipient. The user may also customize correspondence from all available personalization types and tones that the engine has previously created. The personalization engine learns from the choices users make to continually adapt to recipient preferences.

The results:

  • Nova Labs has started beta testing with two dozen companies;
  • The pipeline increases daily with more than 200 companies expressing interest in the product;
  • More than 5,000 emails per week go through the Nova system; and
  • Nova-personalized emails perform more than 500 percent better on average than control (non-personalized) emails based on numerous experiments using real-world campaigns for both prospect engagement and response rates.

Nova Labs is continuing to meet its own internal milestones and is aggressively pursuing new client growth using its proprietary tools, product, and technology.

About the Author

Bryan Pirtle, SDM ’13, is chief technology officer of Nova Labs, Inc. He employed techniques from MIT’s System Design & Management (SDM) program to help shape the core of Nova Labs’ technology strategy and roadmap.