Improving Response to Customer Leads
Gene Linetsky is the Chief Technology Officer for RealtyShares.
In real estate, finding operators looking to raise money is often grueling. At RealtyShares, we realized that optimizing for the quality of each borrower lead and for the fastest time to first contact would help us improve our closing performance and make for a better experience for our potential borrowers.
To accomplish this, we are looking to data. With information obtained from past engagements, we can use machine learning to discern the pertinent features of a lead and potentially determine in real time how to present the right ones to the right sales representatives. In this way, data can help us identify quality opportunities and turn new leads into new clients. The ultimate objective: operators experiencing faster turnaround and investors enjoying a higher volume of better-quality investments.
We all know how important speed is for our borrowers and investors, and being able to respond to qualified sponsors with valuable projects in minutes rather than hours means more potential quality deals for our investors.
Next in the Tank
For Phase I of our plan to improve lead performance, we needed to change the way leads are distributed to our representatives. Rather than doling out leads one-by-one in a round-robin fashion, we now present them to the entire sales team all at once on a new interface, so that anybody available to react to that lead can now pick it up. Each rep ends up contending for new leads, a system our sales pros affectionately dubbed the “Shark Tank.”
As a result, the Shark Tank process (which inadvertently ended up feeling like gamification of lead routing) has energized the team and led to both more camaraderie and some friendly competition. And in the first week Shark Tank has been in place, reaction time has significantly decreased. Operators now get their deals processed faster and more efficiently, resulting in an overall more accessible platform.
Bringing Tech to Residential Debt
Now that leads are addressed more swiftly, we plan to introduce Phase II: machine learning. Ideally, we want to learn from our data (or, more precisely, we want our predictive models to learn from data, and keep learning from it) and see, from the borrowers we previously converted, which ones ended up as customers and which leads took up considerable time, only to be abandoned. The goal is to use machine learning to help determine the relative value of leads in real time, moments before they reach our sales team.
We can also monitor individual performance characteristics — even factored by time patterns — of our reps, and allocate leads according to each representative’s strengths and skills. For instance, if Representative A’s personality is better suited to Midwest properties, while Representative B has more success with coastal regions, we can tailor our lead distribution to match each rep’s forte. Essentially, we want to build the entire residential debt system utilizing all of our pre-existing data — not just narrow lead data, but RealtyShares’ entire history and experience thus far.
Potential borrowers are flooding us with thousands of applications, a rate that allows us to make intelligent determinations. Likewise, we diligently monitor data in all stages, from sourcing to converting. That way, we can cross-reference every single new lead with all the data we’ve accumulated to improve performance for funding, servicing and completing great real estate projects.
At RealtyShares, data is the crown jewel of the company. By using our technology to add value to our residential debt product, we strive to incorporate data into every facet of business operations. Improving lead response is the next step in our mission to give investors and borrowers the best experience possible.
