The other night, I was watching Law & Order SVU when a scene cut to Morales, the analyst. It was funny, because that’s my name too … Well the analyst was briefing the team on critical data he’d pulled from a suspect’s USB key. It was an Excel spreadsheet. But Morales didn’t just dump a bunch of analyst jargon on them. No, Morales told a story. He framed his insights to suit the audience case.
And then it hit me: Why don’t we do this more in business analytics?
Too often, our job looks like navigating lifeless grids of numbers. But data analysis isn’t just about crunching figures—it’s about communication, interaction, and collaboration. Inspired by Morales, I want to explore this in my FBI series of blog posts, starting with the reason for investigating [1], reaching out for clues [2], and investigating per se [3] . Because if an FBI analyst can make data feel like life-or-death intel, surely we can sell a budget forecast or market trend with a little more impact.
1. Is this Case Worth the Badge?
Before launching a full-scale investigation, let’s establish the context. A sudden drop in conversion rates from recent events has set off alarm bells. But is this just a blip on the radar, or are we dealing with a serious crime against the quarterly target? Before we start rounding up suspects, we need to verify if there’s really a case to crack. If the numbers point to a real threat, it’s time to move fast with smart, tactical adjustments. If not, there’s no need to waste resources on a wild goose chase. The badge only comes out when the case is real.
Who’s the victim? A sudden drop in third-party event conversion rates.
Why solve this crime? This drop threatens the quarterly target by $300K; not just a minor setback, but a direct hit to business performance. Now we have a case. But is this just an isolated incident, or part of a bigger plot? Does this drop jeopardize broader strategic objectives, like market expansion for this quarter’s product push? Before launching a full investigation, we need to determine if this case is a priority or just background noise.
What does the police need? The marketing department must drive corrective actions before the quarter ends. We need to provide them with guidelines to quantify team efforts in time for next week’s meeting. They don’t need an overcomplicated post-mortem; they need clear, actionable intelligence to course-correct in time. This involves listening beyond the surface because what stakeholders ask for might not be what they actually need. We also need to separate symptoms from root causes to make sure we start solving the right problem. Indeed, stakeholders might ask for a solution without fully understanding the root cause. Separate needs from expectations. Instead of immediately building a new report, use an RFI (Request for Information) form to structure their request. This ensures you’re solving the real problem, not just responding to surface-level symptoms.
How do we crack the case? Different investigative approaches will reveal different layers of the truth. Write a coronor report to describe what happened? This is the descriptive analysis. Perform a forensic examination on crime scene to explore it and see what stands out? This is the exploratory analysis. Drive a detective investigation to explain why did the crime happend? This is the explanatory analysis. Or share your profiler’s expertise to predict what would be the killer’s next move? This is the predictive analysis.
2. Follow the Chain of Command
Before cracking the case, gather the evidence. This is your data wrangling phase. Hunt for clues across multiple sources and research leads thoroughly. Clean the data, and structure it to ensure high reliability. But how you collect your proof matters. Just like a police officer must obtain depositions properly for them to hold up in court, you need to retrieve your data through the right channels. Get the data the right way, or risk losing your case. However, you can play the right hands of influence to smooth out tough negotiations and keep the information pipeline flowing.
Harness Momentum and Enthusiasm. Be quick, stay sharp, and keep the initial energy going as long as possible. The faster you act, the easier it is to maintain engagement. If you lose interest, so will others.
Leverage the Rule of Reciprocation. To receive favors, you need to start by giving some. So before making big asks (budget, resources, access), offer “fast wins”: quick and small insights that provide immediate value. Sweeten the deal by offering a premade plan when requesting something on a shorter timeline. Throw in the fifty-dollar bill like the police officer does with the dealer. Spend an hour working on something your stakeholder cares about to convince another stakeholder to help you.
Cultivate a Sense of Community. Tie their burden to yours. Reinforce the project’s mission and your best key metric; make them feel like part of something bigger. Build team spirit ahead of your investigation, build a sense of belonging. Share wins and visibility with your informants. “Just noticed that your leads from X campaign are converting 2x less than usual. You might want to double down on that. Hey, you know what? I’m already working on a similar topic. Wanna join forces by sharing your insights?” Finally, turn passive stakeholders into active allies by asking questions that invite collaboration rather than just agreement or disagreement. Instead of asking “Should we investigate third-party events?”, ask “What patterns are you seeing in third-party event performance?”
Play with Ego but Smartly. Use competitiveness to your advantage. Benchmark performance, ask your informant how his team stacks up against others. Nothing fuels motivation like healthy competition. But use peer pressure as a last resort. Never cuff a dealer that can still be an ally down the line. Show restraint. Never criticize their contributions publicly, do it in private.
Secure a Sponsor. Navigating an organization without the right allies is like running an undercover operation without backup; you won’t last long. Instead of spreading your efforts broadly, focus on intensive influence by securing a powerful sponsor. To identify the right one, look for influential figures whose goals align with yours. These aren’t just senior executives but also power players. Those who may not have the highest rank but hold the most sway over decisions. Earn their buy-in in your strategic partnership by showing how your work benefits their objectives, by sharing insights that position them as the drivers of success and by making your wins their wins. Then, instead of pushing your agenda across multiple departments with limited impact, work intensively within your sponsor’s network. They will actively champion your cause, opening doors that would otherwise remain shut. This allows you to shortcut bureaucracy, gain priority access to resources, and move decisions forward faster.
Adapt through Strategic Compromise. To echo what we have introduced in the previous point; at the end of the day, influence isn’t just about persuasion but it’s about power dynamics. Like the streets, or the force, your company isn’t a democracy. You can use all the influence you have, you’ll still have to align with the most influential department and provide them with maximum visibility on strategy, timelines, and priorities. At the end of the day, follow the chain of command.
3. Confidence in your Work is Key
Once you get your clues, it is time to analyse them to retrieve insights. But if you don’t trust your analysis, no one else will. So you need to own your insights by developing a credible analysis. On TV, you’ll see FBI analysts like Penelope Garcia or Spencer Reid in Criminal Minds perfrom fancy coding and statistics. But real data analysis is different. Your job consists in finding meaning in the clues you’ve gathered, and apply that to a specific business context to identify insights, investigate hidden patterns, develop and test hypotheses and predict outcomes. So let’s show some critical thinking and let’s give the police what they need.
Coronor Report, aka Descriptive Analysis. A coroner’s report determines the cause of death based on medical examination. It describes factual findings, such as time of death, injuries, and toxicology results. It documents physical evidence without investigating external factors. Hence, it is similar to descriptive analytics, which summarizes past events without drawing conclusions about causes. The primary goal is to establish what happened by summarizing key metrics and identifying deviations from expected results. This is achieved through data visualization, descriptive statistics, data aggregation, and structured reporting. The most common use case involves KPI tracking and obvious anomaly detection during the weekly performance reviews.
Forensic Examination, the Exploratory Analysis. The forensic examination goes beyond the coroner report to determine how and why an event occurred. It involves exploring the crime scene and identifying what patterns stand out by thoroughly analyzing its evidence, fingerprints, ballistics, and DNA. It seeks to uncover foul play, reconstruct events, and provide leads for criminal investigations. It is similar to exploratory analysis, which investigates relationships and anomalies in data without forming conclusions on the killer’s identity yet. The key objective is to mine data to investigate unexpected trends and patterns in order to segment and cluster audiences, identify opportunities, detect outliers, uncover correlations and potential influencing factors that may not be immediately visible. This analysis is frequently used to further investigate the potential key insights retrieved after the weekly performance reviews.
Detective Investigation, the Explanatory Analysis. The detective’s job is to gather concrete evidence to determine who committed the crime, how it happened and why. It involves connecting the forensic examination with more context by interviewing witnesses, reconstructing events, and following leads. It relies on logic, deductive reasoning, and fact-based analysis. This is similar to explanatory analysis, which explains why something happened and pinpoints the key drivers of change. A typical use case is determining and benchmarking the factors behind performance drops to cross-validate hypotheses. This is done through regression analysis (linear and logistic), A/B testing, inferential statistics, confidence intervals, probability distributions, and causal inference.
Profiler Expertise, the Predictive Analysis. A profiler gets inside the mind of a criminal to understand their motivations, and behavioral patterns in order to identify suspects or to predict the killer’s next move. It uses psychology, past case studies, and crime scene analysis to build a psychological profile of the offender. It is similar to predictive analysis, which forecasts future outcomes based on historical data and trends. In order to develop forecasting models and prioritize opportunities, you’ll need to use methods like machine learning models (supervised learning), time-series forecasting and classification models.
Explore more
For more on how to shape the evidence you’ve just gathered into a story that sticks, read the second post in this Law & Order-inspired series: Crack your Case Like an FBI Analyst: Turn Your Evidence into a Compelling Narrative.