I have been very fortunate to work with some world-class business leaders throughout my career. The biggest thing I’ve learned is that the best leaders operate their business with machine-like precision. Data and metrics are the lifeblood that runs a multi-billion dollar business because it cuts through the noise and identifies the right lever to pull.

The emergence of science in B2B sales
For the last four years, I worked with a team of sales leaders who proudly called themselves ‘data junkies.’ They carefully analyzed fact-based leading indicator activities that led to repeatable success and matched them against things that don’t work. This team designed a scientific model that, if executed, would guarantee success, based on hard data, not gut feel. We kept emphasizing to new sales leaders that the best sports managers know what plays and tactics are more likely to result in a goal, and we expected the same from them. “Moneyball” had been quoted 100s of times at every SKO. If you walk into a sales team’ business review, you would think you mistakenly walked into a math class with the whole room of sales leaders crunching numbers on a piece of test paper.

Make bricks without data
With a growing appetite for using sales activity data to optimize B2B, I built a team of high-caliber, incredibly smart analytics experts. They came from the investment banking world or B2C giants. They had experience modeling human behaviors for financial derivatives or the biggest online retailer on the planet. Building predictive models for an enterprise software company was a new challenge, very difficult to solve, but highly rewarding – since every bit of optimization we could harvest led to a needle-moving revenue impact.

However, the biggest challenge was not training the models – in B2B analytics, you can’t skip the data collection and go straight to the insights. This team built sophisticated models trying to decode sales behaviors, which I was thrilled to see their progress on, yet deeply agonized by it.
Why? Because the inputs were too fragmented, too high level, with too much human intervention and bias. For example, “one” sales meeting can be a 20-minute regular check-in with an analyst, or a full hour deep dive with the VP, directors, and the command chain with the CEO involved. To add more complexity, if you wanted to build the predictive model using the number of engagements, there was usually only one contact manually entered into Salesforce while in reality there were 13 contacts engaged.

I needed to solve this problem. There were tons of vendors trying to sell me variations of solutions that “score” the deal but only used inputs such as opportunity stages change, amount change, deal age. This type of forecasting tool skips the hard thing of generating the right inputs. Whenever I questioned the accuracy of their model, I got a cliche answer: garbage in, garbage out. Tell me something I didn’t know, please!

That got me thinking: in the banking world, one skewed coefficient of the risk model can lead to a market crisis. Similarly, in the sales world, skewed or incomplete activity data would send a false signal to the sales leaders, who are making important decisions about organizational design, capacity planning, forecasting, commission, ecosystem, etc.

As a smart analyst once told me: it’s an analyst’s best dream to unlock business interactions’ data in the enterprise world. Whoever gets there first, will have an unfair advantage.

Right activity data in the right place
This journey is exactly why I joined People.ai. I found a group of innovative and talented people who deeply understand this pain in the B2B enterprise software world, who shared a vision of digitizing the activities of all knowledge workers using AI. We are relentlessly building a product that brings static activity data to life, intelligently matches that activity to the right person/team/business process in the complex data environment of enterprise, and detects dynamic benchmarking for account health, deal engagement, and team productivity.

I believe that the best job in the world is to build something now that will do the work for you someday. Life is short, instead of manually entering activity data into Salesforce, or running regression models using completely wrong inputs, how about we build meaningful connections, solve real business problems and live a great life? AI has the power to do work that frees humans to focus on things AI can’t do — make a human impact in this world.

I am thrilled to be on a team that is building this into reality. People.ai can now collect live data while the critical business interaction is still happening and even better – it has the power to retrieve all the historical activity data since the beginning of the business. There are 18 million sales professional in the U.S, another 10 million sales engineer, customer success, services professionals. And we haven’t even talked about the decision makers — the economic buyers in corporate America. Think about the market opportunity ahead!

In the new world of enterprise selling, sales should have all the business activity history and sales managers should know which account to invest more time on based on the live ROI of the deal, and be able to decide the most optimal amount of effort that should be aligned to each account. Marketers should have an evergreen, rich and live contact database. We are working on it here at People.ai.

Are you one of us who shares this vision? Hit me up if you are ready to be part of something special. We’re hiring!










Dan Zhang is People.ai’s Head of Analytics & Strategy. Prior to joining People.ai, Dan was Director of Sales Analytics & Strategy at AppDynamics. Dan holds dual Master’s degrees from Syracuse University in Statistics and Sociology and a Master of Science in Finance from the University of Illinois at Urbana-Champaign.