A command centre with five dashboards that no one uses is worse than no command centre at all, because it spends real money to manufacture the illusion of analytical capability while the business runs exactly as it did before.
India is in a dashboard boom. Business intelligence licences are cheap, Power BI and Tableau have made visualisation a weekend project rather than a programme, and every large Indian company now has a wall of screens somewhere with a name like “growth cockpit” or “war room.” The energy is understandable: Indian consumer businesses are among the most operationally complex in the world. A Fast-Moving Consumer Goods (FMCG) company sells through millions of outlets across a hierarchy that runs from national headquarters down through branches and territories to the individual distributor and store, and the kirana channel — the overwhelming majority of volume — reports almost nothing back. The hunger for visibility is legitimate. What is not legitimate is the metric by which these programmes quietly get judged: dashboards shipped, sources connected, screens installed. None of these is an outcome. The only outcome metric that matters is time-to-action: how much faster does a manager do something because the system exists?
Across a decade of deployments — as a vendor building these platforms for large Indian enterprises, later as a user of them inside global agencies, and now as an advisor evaluating them — the pattern repeats with almost embarrassing fidelity. The diagnosis is usually correct: the data the leadership needs genuinely exists, scattered across a sales-force automation tool, a distributor management system, a trade promotion system, a Customer Relationship Management (CRM) platform, the Enterprise Resource Planning (ERP) backbone, mobile field apps, and a data warehouse — seven systems in one engagement I remember vividly, each owned by a different function, each reporting on its own calendar. Management sees a consolidated picture once a month, weeks late, by which time the operational window has closed. So far, so real. The proposed cure is a single environment where the same data is visible daily at every level of the hierarchy. Also correct. The failure arrives in two specific places, and almost never anywhere else.
The first failure is the integration estimate. On the architecture slide, seven data sources are seven labelled boxes with arrows. In reality, each box is a different data steward, a different schema, a different update frequency, a different access-control regime, and a different opinion about what “secondary sales” even means. The honest pre-project question is deceptively simple: for each source, what is the update frequency, who owns it, and in what format does it arrive? If that question produces three different answers from three internal stakeholders about the same system — and it usually does — the integration timeline should be doubled on the spot. I have never once seen the integration line in a command-centre budget come in under its estimate, and I have watched it consume the analytics budget whole. The visualisation layer, the part everyone debates in workshops, is by now close to a commodity; the data layer is where the engagement is won or lost.
The second failure is deeper: surfacing broken data beautifully. A well-designed promotion-effectiveness dashboard sitting on top of a distributor system where promotion-level tagging is incomplete does not create insight; it creates confidently formatted wrongness. The real work in that engagement was never the dashboard — it was bridging the data quality problem at source: getting schemes tagged consistently, getting distributor billing data to reconcile with the manufacturer’s records, getting field teams to capture execution honestly. That work is unglamorous, politically awkward, and invisible in a demo, which is why programmes skip it and go straight to the screens. The screens then quietly lose their audience within two quarters, because managers learn faster than steering committees which numbers can be trusted.
This is where the dual lens earns its keep. The brand voice — in this case, the organisational-ambition voice — wants the command centre as a statement: we are a data-driven company, look at the wall. The evidence voice asks one question of every proposed dashboard, and it is the question I now put at the top of every design document: what does the manager do differently on the day they see a declining trend here, versus the day they would have seen the same trend two weeks later in the monthly report? If the answer is concrete — she calls the area sales manager that afternoon instead of waiting for the review cycle — the dashboard has a commercial case, and the case has a value you can estimate. If the answer is “they are better informed,” the dashboard is decoration, and decoration has a carrying cost: licence fees, refresh pipelines, a team maintaining it, and — the expensive part — executives who believe the capability exists and therefore never fund the version that would.
That last cost is why unused dashboards are worse than none. An organisation with no analytics knows it is flying on instinct and behaves with appropriate humility. An organisation with five ignored dashboards believes it has instrumentation, cites it in board decks, and stops asking the question. The illusion of capability crowds out the acquisition of capability.
So, Monday. Take an inventory of every dashboard your organisation pays to keep alive, and for each one ask two questions: who looked at it in the last thirty days, and what decision changed because of it? Retire, without ceremony, every dashboard that fails both — sunsetting is not an admission of failure, it is the recovery of attention. For the survivors and for anything new, require a one-line action contract before any build: the named role, the trigger condition, and the action taken, with time-to-action as the metric reviewed quarterly. And before you approve the next integration estimate, ask the ownership-frequency-format question of each source yourself, in one room, with all the stewards present. Count the contradictory answers. Then double the timeline, or better, spend the first tranche of the budget fixing the data at source — because a smaller, plainer dashboard on trustworthy data will beat a wall of screens on doubtful data every quarter of the year.