Case Studies · Data stories

    When data becomes decision.

    Five real stories — different sectors, same logic: turning thousands of daily signals into concrete, measurable, repeatable decisions. No glossy slides, just problems tackled and numbers that moved.

    Tell us your case
    Case · 01
    Logistics · Deliveries

    On-time deliveries, every time: how we eliminated route delays

    « Before, our best employee spent two hours every morning drawing routes by hand. Now the system does it in 30 seconds, and deliveries are on time. »
    The problem

    A local delivery service with 15 vans planned routes manually or with standard maps. Delays were frequent, fuel costs kept climbing and customers called to check where their goods were.

    What we did

    We built an intelligent routing platform that computes optimal paths in real time — factoring traffic, customer priority, time windows and vehicle capacity. Drivers use a simple app, customers get automatic tracking notifications.

    Outcome
    −34%
    average delivery time
    −28%
    km driven per day
    +92%
    customers reordering
    Case · 02
    HR · People Analytics

    Evaluating 280 people without bias: the framework that changed promotions

    « Calibration meetings went from 3 days to half a day. And no one leaves feeling unseen anymore. »
    The problem

    A B2B services group ran subjective annual reviews: manager perceptions, poor traceability, internal tensions at every promotion cycle.

    What we did

    We unified data from CRM, ticketing, project management and calendars. The result is a People Analytics cockpit with output, collaboration, customer impact and growth metrics — weighted by role and seniority. Managers interpret, data measures.

    Outcome
    −63%
    internal review appeals
    +22%
    top-performer retention
    100%
    promotions with objective evidence
    Case · 03
    Retail · E-commerce

    Reading the market 6 weeks earlier: how we caught the weak signals

    « We stopped reacting to sales and started anticipating the market. »
    The problem

    A mid-market fashion brand discovered sales trends 4–6 weeks late. Result: stock-outs on best sellers, aggressive discounts on the unsold.

    What we did

    We built an early-warning model that correlates on-site searches, out-of-stock rate, cluster micro-trends and external signals. Buyers receive weekly alerts on categories to reinforce or reduce, weeks before they show up in sales.

    Outcome
    −42%
    unsold at end of season
    +19%
    full-price margin
    6 wk
    lead-time gained on the market
    Case · 04
    Performance Management · Productivity

    Clear goals, measured productivity: the cockpit that aligns teams and outcomes

    « People always know where they stand versus their goals. Managers coach, they don't judge after the fact. »
    The problem

    A professional services firm set yearly goals on paper, with no continuous tracking. End-of-year surprises: wrong tasks, overload, under-utilisation. Fairness at its lowest.

    What we did

    We built a performance management system crossing individual and team goals with real operating data: billable hours, deliverables, perceived quality, project contribution. Every person owns a real-time personal cockpit; managers see the team, leadership sees productivity per business unit.

    Outcome
    +31%
    measured productivity
    +48%
    goals hit on time
    Real-time
    continuous feedback
    Case · 05
    Real Estate · Territory

    City-scale property census: giving every building a face

    « We used to knock at random. Now we know who to talk to, when, and with what offer in hand. »
    The problem

    An operator with urban-scale ambitions had no real map of the stock: vacant buildings, fragmented ownership, off-market opportunities hidden in cadastral files.

    What we did

    We built a property census integrating cadastre, company registrations, ownership profiles, historical listings and market signals. Every building became a living record: who owns what, what type of owner, likelihood to sell, rent or redevelop.

    Outcome
    100%
    buildings censused
    3.2×
    off-market opportunities captured
    −70%
    scouting time per new project