People Analytics – what is it?
A quick Google search tells us that People Analytics “refers to the method of analytics that can help managers and executives make decisions about their employees or workforce”. This semester, I took a course on People Analytics with beloved HBS professor Jeffrey Polzer. We covered various cases about how companies are utilizing data and algorithms to make better decisions about people. For example, Google’s Project Chameleon allowed the gTech group at Google to better match employees with internal job openings via an open market staffing algorithm. Employees and managers would preference each other, and via the Gale-Shipley Deferred Acceptance Algorithm, would be matched with their optimal role. The algorithm guaranteed employees would get top choices and discouraged any gaming. This approach to staffing a very dynamic function at Google eliminated the need for a centralized staffing function and / or an external consulting project. 
In our People Analytics course, the examples of organizations using big data and analytics to support their HR and people function were diverse and numerous. Companies are using data and analytics to inform issues such as employee hiring, firing, and retention. More specific use cases include measuring salespeople’s performance, or helping patent examiners at the USPTO more quickly and flexibly review patent applications (a machine learning algorithm improved the Patent Office’s internal search capabilities, enabling patent examiners to work from home and search for prior art on their laptops). Even more interestingly, organizations are formally codifying these practices; in a recent Deloitte study, 69% of companies surveyed reported that they are integrating data from disparate sources into a people analytics database that can be analyzed.
Why should consultancies be concerned?
Prior to coming to Harvard Business School, I worked at the Boston Consulting Group (BCG) for 3 years; the people analytics use cases I mentioned in the prior section made me concerned for the future of a large portion of BCG’s business. There are two major categories of projects I would be worried about if I were a partner at BCG: 1) organizational transformation, and 2) performance management. In the sections below, I provide a more detailed description of each type of project, based on personal experience.
One of my most fulfilling and successful projects at BCG was an organizational redesign at a large consumer packaged goods company. The company had not been performing well, and so the CEO mandated an organizational restructuring and layoffs. The CEO hired BCG to help facilitate the redesign process. We used prior expertise to help the company: 1) define the organizational design principles to adhere to (e.g., “no manager will have fewer than five direct reports”, “we will always optimize for fast decision-making”, etc.), 2) conduct workshops to make design and staffing decisions and 3) track the organizational and financial data to ensure the effort was on track.
BCG helped the company hit its organizational and financial goals, but aided with machine learning and a people analytics database, in the future the company could automate much of what we did. For example, armed with historical functional performance, the company could build an algorithm to predict what type of organizational structure enables the most success for marketing (e.g., with 8 direct reports and 3 layers below a manager, a marketing function increases successful product launches by x%). This type of algorithm could inform design decisions. Additionally, algorithms could inform staffing decisions by correlating individual traits with job performance (e.g., for managers of function X, 5 years of experience in function Y is particularly useful, etc.). In the future, I envision a world where large organizational redesigns are supported by algorithms, not consultants.
A second and very common type of project I worked on at BCG is what I will refer to as “performance management”. A pharmaceutical manufacturer had hired us to de-bottleneck a quality control (QC) laboratory. The QC lab was receiving sample product to test at a frenetic rate, and the testing requirements for each sample were cumbersome. The result was a backlog of hundreds of product samples. To address the issue, we recommended a few process changes, and then implemented a tracking system to ensure product testing was occurring at an acceptable rate. In the future, though, I would imagine a people analytics database will enable automated tracking. For example, imagine when an employee swipes in to work, a timer starts. Then, when an employee wants to test a sample, he or she must swipe his or her card again, and a separate timer starts tracking test time. These types of data would enable the large pharmaceutical manufacturer to track performance in real-time, versus hiring a consultant to come in and diagnose / remedy the problem.
How is BCG responding? What else needs to be done?
In response to the threat of AI and big data / analytics, BCG has created multiple product groups to support consulting teams and clients. For example, the BCG Gamma team provides support to consultants via big data and analytic tools. The team specializes in AI and is equipped with data scientists. Another example of a digitally-focused investment is BCG’s Digital Ventures (DV); DV helps companies define a digital strategy, and also has the expertise (software developers, etc.) to help companies implement digital strategies. DV even has raised a venture fund to incubate seed stage ventures on companies’ behalf.
BCG’s approach here has been to continue to invest in expertise that will create value for companies as they embark on digital journeys. Said differently, BCG is OK with companies building people analytics expertise that will eventually make certain consulting projects obsolete, as long as BCG can find ways to capture value elsewhere on that digital journey.
My worry for BCG is that they will not move quickly enough to create value for clients in the people analytics space. As we saw in our course work, when algorithms are at work, firms create a competitive advantage through learning effects. In other words, if I can build algorithms before you, mine will get smarter and continue to maintain an advantage over yours as long as I have quality data to feed the algorithm. Creating value via learning effects by rapidly feeding algorithms data is counter to BCG’s operating model; the firm only engages clients when asked, and is constrained by the pace of clients’ innovation. A firm like Google or Amazon, which is ferociously feeding its algorithms data through platforms such as Alexa, is much better positioned to create and support an algorithm to help companies solve hard people analytics problems. If these firms are able to build strong and very smart algorithms, the algorithms could render services like BCG Gamma or DV obsolete.
My recommendation to BCG is to be more open with their intellectual IP. Create online platforms that showcase algorithms that teams such as BCG Gamma have built, and allow companies to use these platforms on their own data sets. This will open up the algorithms to more data and enhance learning. If a company does not have the expertise or data to feed the algorithm, this represents another revenue opportunity for BCG (i.e., BCG can charge the company for a project to gather data and feed the algorithm).
BCG has started to do this through their People Advantage Platform. However, in current form the platform only offers clients benchmarks, not predictive capabilities. If BCG wants to win the algorithm battle, the firm would be wise to offer predictive capabilities as well. As I mentioned, I would expect resistance to this approach given how protective BCG is of its intellectual IP. For this reason, I would keep this more open version of the platform as a separate group, and potentially even rebrand it so that it is not viewed under the BCG umbrella. This would allow the platform some shelter from the typical operating processes / IP rules at BCG.
 “Matching Markets for Googlers”, Cowgill and Koning, HBS Case #N1-718-487
 “The Future of Patent Examination at the USPTO”, Khanna Mehta and Choudhury, HBS Case #9-617-027