My first people analytics job was unprecedented and completely invented. It yielded adventure, fulfilment and lessons that I thought might be useful for anyone engaged in people analytics.

I had been hired as a data consultant–a brand new occupation–at a leading HR advisory firm from a quasi-academic role at an international organization, conducting global labor market research. I managed to secure the consulting role by writing a business plan to assuage the hiring manager’s fear that I might not relate to clients or deign to communicate via bullet points.

My first assignment was to monetize the disparate compensation survey assets run by dozens of independent offices around the world. In exchange for access to a global database for their local clients, I convinced the survey owners to surrender their surveys’ meta-data–participants, compensation elements, metrics, etc. I cataloged the information to construct a global, indexed, searchable database that I distributed via CD ROM.

I was able to offer compensation consulting services to global clients who could now more easily and quickly examine their workforce’s global pay competitiveness. I earned a promotion and began my new professional incarnation as an executive compensation consultant. But proxy statements were boring, and the repetitive work was unfulfilling. I yearned for greener pastures. If there was one thing I learned in compensation consulting, it was that the money was on Wall Street.

Through a colleague I wrangled a meeting with the head of compensation at a Wall Street investment banking firm. He took a look at my eclectic background and offered me a job, even though he did not have an open position. The role–Senior Compensation Specialist–was a step down but paid significantly more than my consulting job. My prospective manager was an organization-savvy, well-read, interesting conversationalist and open to new ideas. I accepted his offer.

Ever since, I have been an advocate of hiring for potential–looking beyond the current open roles to imagine what a candidate with unique skills and aptitudes might bring to the organization–and making room for them. The first job fit might not be perfect, but the next one can be a home run. It’s important not just to identify, but also to nurture such talent.

As promised, in addition to the bread and butter compensation tasks of job matching and survey submissions, my manager threw me into a variety of “human capital” related projects. We were in a “personnel” world, but he was open to new paradigm of employees being viewed as an investment. Projects ranged from figuring out housing new reimbursement parameters for Hong Kong staff in light of changes in local tax regulations to determining what attributes made the best Private Client Services regional manager. The pace and difficulty were exhausting, but the variety was exhilarating.

I’ve made it a point to always mix things up for my team so that they are exposed to different stakeholders and different business economics. The variety of situations and challenges spurs innovation as they make associations and connections across their experiences. HR budgets are invariably tight so the resource constraints force creativity and strategic choices around what to work on. Rather than complaining about insufficient budget, deliver results that spur demand and investment. Look beyond HR for sponsorship if necessary.

Despite being a major Wall Street firm, the organization’s Human Resources Information System (HRIS) was a Paradox relational database (it was upgraded to PeopleSoft years later, my first exposure to the joy of ERP implementation). The HRIS system and the data were fragile and incomplete. In order to evaluate the Analyst and Associate program (Analysts are hired right out of college and work for 2-3 years before leaving to earn MBAs and return as Associates), I had to find resumes for hundreds of people going back five years and distill all the information in them to variables for use in modeling. I also had to teach myself Paradox.

I am always surprised, often dismayed, but never deterred, by leading organizations’ poor HR systems and data infrastructure. HR systems and processes were designed in a pre–big-data world. To be an effective people analytics leader, you have to “get your hands dirty” with your organization’s data, processes and technology. You need to understand the data-to-metric pipeline in order to represent results with confidence.

One of the first things I look at was attrition which was a concern for management. I imported vast tracts of data from Paradox into a massive spreadsheet and crunched out attrition rates from every conceivable angle. I used Stata, a statistics package, to run survival analyses, Cox proportional hazard models and logistic regressions to understand the dynamics of attrition, including variation across gender and ethnicity. The final report reduced complex models into simple tables and charts, with an up-front, one-page editorial executive summary.

An HR leader showed their business leader the executive summary and suddenly the attrition figures started to be referenced in meetings around the company. A quarterly report worked its way up the management chain and in short order I received a summons to present the next quarter’s report at the 14-member Operating Committee that ran the company. With some trepidation–remember, I was a recent arrival in the commercial world, Wall Street and HR–I arrived at the Board Room in new shoes and my lucky Metropolitan Museum of Art golden elephant cuff-links. My manager received an invitation, too, and sat across the room to gesture pre-arranged signals, including the frequent “slow down and let them digest the information.” A rich discussion ensued and I was added to the agenda.

Presenting in the board room to executive leadership for the first time is quite daunting. I had my boss’s complete support and he was thoroughly invested in my success. The exposure to leaders early in my career was a blessing. Through those interactions I developed a good sense of how leaders consumed data and what workforce information they needed for decisions that kept their strategy execution on track. In every company I subsequently worked for, I would give my team members opportunities to present to senior leaders. When they are ready and have your confidence, they will make you proud. They become known quantities in their own right and additional opportunities open up for them.

One day, my manager revealed that the Investment Banking Division (IBD) had been trying unsuccessfully to estimate how many Analysts to hire each year. A poor estimate leads to sudden staff expansions or contractions. These unplanned actions incur unanticipated labor cost and adversely impact the firm’s reputation in the labor market. IBD had deputed some Associates to work on the problem; they hadn’t even considered reaching out to HR. My manager felt we had built up enough credibility capital to make a play for the work.

A new head of strategy had just joined the firm and had convened a leadership meeting. My manager said: “Let’s just show up. The head of strategy will assume we are part of the leadership team and the leadership team will assume their new leader has invited us to the meeting.” It worked! When the problem was discussed and the IBD Associates had no solution, we bid for and got the work. And we got to keep attending the meetings and offer input on workforce strategy.

Different organizations might react to this risky gambit very differently. I’ve seen in another organization a senior leader censured for attending a senior conclave without being invited. You must be mindful of your organization’s culture. But wherever possible–and there might be professional risk to you–it is worthwhile to insert yourself into the appropriate conversation to stand up for your team and to win them interesting work.

The IBD Analyst Hiring Prediction Model was no easy thing. Analysts were just one layer in the IBD pyramid, after which came Associates, VPs, SVPs and MDs, organized across sectors and deal teams. To predict Analyst demand, you have to understand the demand profile up the pyramid and the dynamics of recruiting, promotion and termination at each layer. I recalled a mathematical device from my electrical engineering training–the Markov chain–that dealt with transitions across states. What if I modeled each layer in the pyramid as a state, tracked down historical movement patterns, estimated transition probabilities and ran different scenarios–e.g., altering the shape or size of the pyramid or perturbing some other parameter–to see what would happen?

It wasn’t easy to collect the historical data. They were not recorded in any system. I had to go to line managers up and down the chain to work out hires, promotions and exits for the last five years. Of course, I had to explain what I was doing and whom I was doing it for–the co-heads of IBD and the head of strategy). When the day came to present my model to these leaders, a half-dozen of the line managers were also present expecting to be part of the presentation and getting rare visibility to their top leadership. There was enough glory to go around. The model was very well received and even after I left the firm, I received calls asking how to manipulate the mega-spreadsheet underlying the model. This Markov-based internal mobility model was developed independently but contemporaneously with Mercer’s impressive Internal Labor Market (ILM) model.

Hiring people with deep, unique and complementary skills has been a key aspect of my team-building strategy. Different areas of study and field of endeavor bring different paradigms and tools to approach and solve problems. The other learning from this incident was to be proactive and explicit in recognizing everyone who had a hand in the solution. It’s not always about who delivers the final presentation.

As I continued my work at the firm, I got increasingly familiar with the “inefficiencies” rampant across programs and processes around recruitment, diversity, talent management, rewards and recognition. We needed a better approach, a data-driven approach. We needed to apply math, statistics and organizational science to the practice of human resources. We needed to get better at workforce planning. We also needed to arm HR business partners with the right information so they could be true “business partners.”

I came up with a proposal to my manager. Let’s start a new group within his compensation team, but with a remit across all of HR. We can name the team “HR Strategy and Analysis.” Initially it would be a one-man show– just me. We would work directly with the head of strategy (hence “strategy” in the title”) and our work would be dictated by both business and HR strategies. We would save the company money by optimizing labor cost and we would make the HR team more effective. We might be able to get other HR colleagues invited to the leadership team meetings of their respective divisions.

Having leadership buy-in and support is vital, especially for early-stage people analytics teams. When exploring people analytics management opportunities, be sure to identify who your supporters (beyond your direct manager) will be and who will decide on funding the team. Even when people analytics emerges to meet an urgent demand, like it did in this investment banking organization, support and sponsorship can wither away just as easily as it emerged when it suited everyone.

My manager warmed to the idea and secured the required approvals. I was appointed Manager, HR Strategy and Analysis. I was happy as a clam. We developed models to predict attrition, promotion and overall success in the firm to identify the underlying dynamics and factors. We built “people plans” for different divisions. We also worked with qualitative data from interviews and focus groups, using the rudimentary text-analytics tools available at the time, bootstrapping as necessary to measure sentiment and engagement. I eventually hired an analyst from the Public Finance division and we were officially a “people analytics team.”

This is the origin story of many early people analytics teams. Someone in the organization catalyzes “people analytics” and it catches the imagination of strategically minded leaders. The early years are often a struggle as teams hunt for the right projects, work hard to deliver results and make a case for investment. This early career experience helped me take two Fortune 10 companies’ people analytics start-ups–which I populated with the very best talent in the world–through exponential growth and success.