How Smart Services Can Scale Up Your Small Business

MAY 25, 2016 • Blog Post • BY QUARTZ CREATIVE


  • Startups must learn to transition from high-speed hackathons to the slow, methodical process of finding product-market fit
  • SocialCapital uses HPE’s machine learning platform to fine-tune its product and find growth opportunities

One startup's journey from hackathon to growing business

Entrepreneurship is in flux. The spread of hackathon-style events and “project sprints” has created an explosion of new business ideas across industries. Despite the flood, the pipeline for developing nascent businesses is broken.

Whether you’re developing a new line of business inside a corporation or building the next Silicon Valley mega-startup, it can be hard to transition from the high-speed, high-intensity environment of a hackathon to the slow, methodical process of finding product-market fit. Hackathon judges sometimes pick “winners” with great stories but not great business plans, making it even more confusing to tell which ideas have potential.

  • It's hard to transition from a high-speed hackathon to the process of finding product-market fit.

Experienced entrepreneurs know that the customer is the only judge that really matters. Tope Alabi is CEO of SocialCapital, a startup that uses machine learning to bring powerful marketing automation to small businesses. Alabi and his co-founder Obed Eugene are both under 30, and are both the children of entrepreneurs. After a stint in the AngelHack accelerator, Alabi and Eugene entered the Techstars incubator in late 2015.

The pair had developed a psycholinguistic algorithm that analyzes public online communication, and they figured HR departments could use the tool in their efforts to hire people based on cultural fit. “Simply put, we are able to take an email input, go out onto the Internet, find the person's online footprint and turn that into a personality type,” says Alabi.

But their original hypothesis about recruiters didn’t pan out, so the pair iterated and tested until they hit upon a huge discovery: marketers were interested—very interested—in using what they’d built to profile customers.

Their service, which cuts customer acquisition costs and surfaces powerful audience data, provides businesses with “a whole new level of customer intelligence,” says Alabi.

There was just one problem with their discovery: machine learning is complex work and SocialCapital is a team of just two. Instead of taking a small business loan or a big venture investment to explore the market opportunity they’d found, they partnered with Hewlett Packard Enterprise (HPE) to use machine learning as a service, building out their value proposition without expanding the rest of the business.

HPE’s Haven OnDemand is a suite of APIs that gives SocialCapital access to machine learning as a service, which they use to perform sentiment analysis on public social network data from Twitter.

“We use the HPE Concept Extraction API to identify interests, so we're able to identify entities and concepts that people talk about socially,” says Alabi, who worked on voice recognition algorithms at DARPA prior to founding his company. (Dr. Michal Kosinski, a Stanford computer scientist known for his work on individual behavioral differences, is a SocialCapital advisor.)

“Based on the adjectives and functional words that a person uses, we're able to identify the personality type according to pre-existing standards like the Myers-Briggs test,” says Alabi.

The system sorts customers into four categories according to their predicted spending habits: strategic, charitable, compulsive or tight-fisted. Their data also allows them to predict the customer’s motivation along an 11-point rubric.

Combining that data with their interest graph is critical for marketers who have little information about their customer base. “We help them get insight into who their customers actually are,” says Alabi. “It’s like a graph of personality.”

Once SocialCapital’s technology segments a group of customers into buckets, it feeds the customer matrix back to marketers, who can create new ad campaigns that target the customers in each bucket with something that will resonate. “For example, if you were to have an extrovert, you would offer them a discount in a way that allows them to invite more people,” says Alabi.

For entrepreneurs across industries, SocialCapital could be a model business for both tech startups and more traditional SMBs. Lavish corporate hackathons and ample venture capital have receded over the last year, leaving entrepreneurs—even those in tech—to bootstrap and grow sustainably.

  • SMBs use intelligent machines to achieve huge scale without a big marketing push.

By leveraging HPE’s Haven OnDemand platform, the SocialCapital founders were able to avoid the distractions and dilutions of fundraising and the stress of trying to build their own sentiment analysis system from scratch. At its current pace, Alabi says the company will be able to fully afford the HPE services in about one to two years. Until then, HPE provides the service gratis. “That helped us get started on our startup journey, to sort of go from zero to a viable product,” says Alabi.

If SocialCapital succeeds at its mission to bring the power of machine learning to marketers, it will be part of a new generation of SMBs: businesses that use intelligent machines to achieve huge scale without financing a big expensive marketing push, or hiring dozens of engineers before the product is ready.

“We imagine SocialCapital to be the strongest personality artificial intelligence machine in the world” says Alabi. “The applications are endless.”