The analyst was drowning. Forty-seven recurring reports. Every Monday, Wednesday, and Friday, the same ritual: pull data from three systems, paste into Excel, apply the formatting, email to the distribution list. Some reports had been running for years. No one could remember who originally requested them or why.

When she finally asked the recipients if they still needed them, the answers were revealing. “I delete most of these.” “I glance at it occasionally.” “What report?” One VP admitted he’d built his own spreadsheet because the official report never had quite what he needed.

This organization didn’t have a reporting capability. It had a reporting burden: a legacy of accumulated requests that consumed analyst time without delivering value. And this is far more common than most organizations realize.

"How did we end up with so many reports?"

The accumulation is gradual and predictable.

Someone asks for a report. It gets built. The need passes, but the report continues, because stopping a report requires a decision, while continuing it requires nothing. Another request comes in. It gets added. The report library grows.

Meanwhile, no one audits what exists. Reports that served a purpose three years ago still run today, even though the business question they answered is no longer relevant, the person who requested them has left, or the data they contained is now available elsewhere.

Each report also spawns variants. “Can you add this column?” “Can I get a version with different filters?” “Can we split this by region?” Requests that seem small create maintenance obligations that compound over time. One report becomes three becomes seven.

The burden grows invisibly until someone counts. And then organizations discover they have dozens, sometimes hundreds, of reports, far more than anyone is actually using.

"How do we know which reports matter?"

Start with usage data, if you have it. Most BI platforms track who opens reports and how often. The results are usually sobering: a handful of reports get heavy use; most get little or none.

But usage alone doesn’t tell the whole story. A report that’s opened weekly might be essential, or it might be opened out of habit and immediately closed. A report that’s rarely opened might be critical for the one person who uses it for month-end close.

Ask the consumers directly. For each recurring report, find out:

Does anyone actually use this? Not “is it useful in theory.” Does anyone make decisions based on it?

What question does it answer? If no one can articulate the question, the report probably doesn’t answer anything important.

Could you get this information another way? Sometimes reports persist because people don’t know the data is available in a dashboard, a self-service tool, or another report they already receive.

What would happen if it stopped? If the honest answer is “probably nothing,” that tells you something.

The audit process is uncomfortable. People resist retiring reports because it feels like admitting the effort to create them was wasted. But the effort is already sunk. The question is whether to keep spending effort maintaining something no one uses.

"What makes a report worth keeping?"

Reports that survive the audit share certain characteristics:

They answer a specific, recurring question. Not “here’s some data about sales” but “are we on track to hit quarterly targets, and if not, where are we falling short?” The question is clear, the answer matters, and the question recurs frequently enough to justify an automated answer.

Someone acts on them. A report that gets read but doesn’t influence decisions is entertainment, not information. Good reports connect to actions: pricing changes, resource allocation, customer outreach, operational adjustments.

They’re the most efficient way to deliver the answer. If the same information is available in a dashboard, a self-service query, or a system someone already uses, a separate report may not be necessary. Reports make sense for push delivery (information that needs to reach people proactively), less so for information people can pull when they need it.

They’re trusted. A report no one believes is worse than no report. If recipients routinely verify report data against their own sources, the report isn’t serving its purpose. It’s creating extra work.

"Why do people build shadow spreadsheets?"

Here’s a frustrating pattern: an analyst builds a sophisticated Power BI dashboard. It has everything: filters, drill-downs, visualizations, real-time data. The stakeholder sees it, seems impressed, and then asks: “This is great. Can I export this to Excel?”

Two weeks later, that stakeholder is maintaining their own spreadsheet, pulling data manually, doing calculations that already exist in the dashboard. The official report and the shadow spreadsheet diverge. Meetings become debates about whose numbers are right.

This happens constantly, and it’s rarely stubbornness. People build shadow spreadsheets because:

The dashboard doesn’t quite answer their question. It’s close, but they need to combine it with other data, apply their own calculations, or slice it a way the dashboard doesn’t support. Excel is familiar and flexible; the dashboard is polished but rigid.

They don’t trust data they can’t touch. Some people need to see the raw numbers, manipulate them, verify the calculations. A visualization feels like a black box. A spreadsheet feels like control.

They need to annotate and adjust. Real business analysis often requires adding context the system doesn’t have: adjustments for one-time events, notes about what happened, corrections the source data doesn’t reflect.

The dashboard isn’t available when they need it. Offline access, slow loading, system downtime: if the dashboard isn’t reliable, people build backups.

Shadow spreadsheets aren’t the problem; they’re a symptom. They signal that the official reporting doesn’t fully meet user needs. Rather than fighting the spreadsheets, figure out what they provide that the official reports don’t, and close the gap.

"How do we stop the accumulation from happening again?"

The audit clears the backlog. But without process changes, the backlog will return. A few practices help:

Require a business case for new reports. Not a formal process, just explicit answers to basic questions. What decision will this report support? Who will use it and how often? How will we know if it’s working? Articulating purpose upfront prevents “just in case” reports from being created.

Build sunset dates in. Every new recurring report should have an expiration date, perhaps six months out. When the date arrives, the report stops unless someone explicitly renews it. This flips the default from “continues forever” to “ends unless still needed.”

Consolidate before adding. When someone requests new information, first ask whether it could be added to an existing report or dashboard rather than creating something new. Fewer, richer reports are easier to maintain than many narrow ones.

Track the portfolio. Maintain visibility into how many reports exist, who uses them, and what they cost to maintain. Regular portfolio reviews, quarterly or biannually, catch accumulation before it becomes overwhelming.

Invest in self-service. Many report requests stem from people who can’t get data themselves. Enabling self-service for common questions reduces the demand for custom reports. The analyst who would have built the report can instead help users answer their own questions.

"What should our analysts be doing instead?"

When analysts escape the report factory, they can do higher-value work:

Answering questions that don’t fit templates. Novel questions that require investigation, judgment, and synthesis, not just pulling data into a predetermined format.

Finding insights no one asked for. Proactive analysis that surfaces opportunities or problems before someone knows to ask. This is where analytics creates unexpected value.

Improving data quality. Investigating discrepancies, fixing upstream problems, building better data foundations. Work that makes all reporting more reliable.

Enabling self-service. Building dashboards, training users, documenting data sources. Work that reduces future report requests by helping others answer their own questions.

The reporting burden is an opportunity cost. Every hour spent maintaining reports no one uses is an hour not spent on analysis that could change how the business operates.

"Where do we start?"

Start with the audit. Count your reports. Check their usage. Ask their consumers. Most organizations that do this honestly discover they can retire 30-50% of their reporting portfolio with no business impact.

Then fix the process. Build in sunset dates. Require purpose statements. Consolidate where possible. Make retirement normal rather than exceptional.

Finally, redirect the capacity. As the burden lifts, point analysts toward the work that creates more value. The goal isn’t fewer reports; it’s more insight.