Focus on the most impactful applications of predictive analytics

Focus on the most impactful applications of predictive analytics – otherwise the juice may not be worth the squeeze

            In 2016, Deloitte announced the launch of ConnectMe – and HR delivery platform that, using predictive analytics, will personalize and target information made available to an employee. For example, should it become obvious that an employee recently had a baby, the information targeted to he or she on ConnectMe would be about adding a dependent. The stated intent is to increase employee engagement through seamless information access and decrease the cost of transactional work for HR professionals[1]. I’m skeptical of the product for a couple reasons outlined below.

            One con is the perception of lack of privacy based on the data mined and the information shared across platforms. The most obvious potential pitfall is health data. Let’s take a hypothetical case: say Mary is diagnosed with a chronic illness that may impact her ability to perform at work. She may have not yet shared this information with her managers or mentors – and she presumes information indicating her diagnosis (such as specialist referrals or appointments) is sitting securely in her insurance portal. Should “targeted” information on medical leave begin to show up on her portal, she may lose trust in her personal privacy and sensitive data held by HR.

            The second con is an unclear link between ease of access to HR information and employee engagement. It’s widely recognized that employee engagement has a positive impact on financial and organizational performance[2]. However, Gallup categorizes employees as engaged based on ratings of workplace elements, such as encouragement by mentors and meaningful work. I’m skeptical that ease of access to HR information has as statistically significant – or as large of an impact – on employee engagement as other factors such as recognition and opportunities for autonomy in work.

            Given these cons, I’m doubtful of whether or not the incremental effort of integrating back end HR data systems and applying predictive analytics – all while navigating the potential complexities of employee privacy – is worth the effort. Further, the last mention of ConnectMe on the web is circa 2020, and the link on Deloitte’s website no longer works (, indicating the product is no longer offered. There may be a potential upside in efficiencies through streamlining the HR functions – freeing up HR professionals’ time to pursue more strategic projects rather than answering run of the mill questions from employees. There may even be an opportunity to use the infrastructure of ConnectMe – i.e., the algorithms and back end data infrastructure – for more directly impactful applications, such as predictive analytics on attrition or workforce allocation. However, as the product stands, I’m a skeptic. I think that if companies are to go through the effort of implementing predictive analytics across their people data, they should focus instead on the most impactful applications to make the investment worthwhile for the business.

[1] Eisenberg, A. (2016). Deloitte Launches Predictive HR Platform to Engage Workers. Employee Benefit News.

[2] Adkins, A. (2016). U.S. Employee Engagement Reaches New High in March. Gallup News.


Data can build better businesses.


“Big Data” or “Big Brother”?

Student comments on Focus on the most impactful applications of predictive analytics

  1. Wow, I’m really glad to hear ConnectMe doesn’t seem to exist anymore because I completely agree with your points – the risks definitely outweigh the potential benefits, in my opinion. We get this enough from targeted ads and marketing tactics. While I understand the intention and can appreciate the sentiment to an extent, it also feels highly inappropriate and crosses several boundaries. I feel like personal information at that level should either be voluntarily and directly shared by the employee or solicited very transparently with a clear intention. We all know humans are susceptible to biased behavior and also underestimate the power of these biases at the same time, so how do we know that news of a pregnancy won’t subconsciously cause a supervisor to find other reasons why that parent shouldn’t be up for a promotion or change health insurance options knowing an employee might have significant medical care? Really interesting topic, Mary – thank you!

  2. Totally agree with you on this. The privacy concerns are obvious and will pose a big barrier to adoption. But even if that wasn’t an issue, it is difficult to see what the meaningful insights ConnectMe can pick up that will actually drive the outcome desired – employee engagement / satisfaction. Overall, this feels very much like a cool, shiny toy for HR and at best a “nice-to-have”. Definitely not worth the cost of development, implementation and lost employee trust from privacy concerns. Predictive analytics can be powerful and they may be onto something here but they need to identify a key use case / linkage to the desired outcome before pushing out broadly.

  3. Hi Mary Helen,
    Thanks for sharing—I had never heard of ConnectMe prior to your post! I appreciate your skepticism of the platform—it appears to be well-founded based on the two cons that you have outlined. In regard to your first con, data privacy is definitely a hot topic within the field of data science. I am currently taking a course with two of our other classmates, Nico and Jeff, on differential privacy—a rigorous mathematical definition of privacy that, when satisfied, provides a formal guarantee that individual-level information about the participants included in a dataset is not leaked with statistical releases. There is some very interesting research going on over at Harvard SEAS within this space if you are interested (! In regard to your second con, I too do not see ease of access to HR information as being a major driver of employee engagement. It seems to me that a well-implemented mentorship program would be worth more investment in resources than this HR delivery platform if increased employee engagement is the ultimate goal. Overall, I am on the same side as you—any example of predictive modeling is interesting to me, but perhaps there are more impactful examples than ConnectMe.

  4. Just to add on, it reminded me of this quote from Laszlo Bock interview: “In my mind, there is a very clean line between personal data in your home and personal life and personal data at work, and you should never cross that line—unless you’re asked by an individual to marry those data sets, and their company approves and they approve.”

  5. I think your skepticism is spot on! The last thing employees are clamoring for is more targeted HR in their personal business. I think this is a great example of well-intentioned People Analytics creating a “Big Brother” feeling that alienates rather than empowers employees. If I want to add a dependent, I will. Making HR functionalities simple and straightforward but still up to the employee seems like a much better avenue – focus the People Analytics team elsewhere on things that can be more approachable and engaging like high-fidelity mentorship matching or a more systematic and better informed performance review process.

Leave a comment