8 Data Quality Issues That Quietly Break Your Reports
Eight data quality issues that quietly break business reports, from duplicates to mismatched definitions, with how to detect each one and fix it for good.
· 4 min read · Summarix team
The most damaging data quality issues are not the obvious ones that crash a spreadsheet. They are the quiet ones that produce a believable number that is wrong: revenue double-counted by a duplicate sync, a region missing because of a spelling variant, a month short because of a time zone. Here are eight that turn up again and again in business reports, how to detect each, and how to stop it recurring.
1. Duplicate records
What happens: the same order, invoice or customer appears twice because an export was appended twice, an integration retried, or a record was captured in two systems. Totals are inflated, often by a small, plausible amount.
Detect: count rows per unique key (invoice number, order ID). Any key with a count above 1 needs a look. Prevent: enforce unique keys at source and deduplicate on import.
2. Inconsistent categories
What happens: ‘Cape Town’, ‘CPT’ and ‘Cape town ’ become three separate branches. The chart shows your biggest branch as mid-sized, and nobody notices because each variant looks reasonable.
Detect: list the unique values of each category column and scan them. Prevent: use dropdowns or lookup tables instead of free text for categories.
3. Missing values treated as zero (or ignored)
What happens: blank amounts become zero and drag the average down, or blank regions are silently dropped from a regional breakdown, so the parts no longer add up to the total.
Detect: report the count of blanks per column, and check that category subtotals sum to the grand total. Prevent: make key fields mandatory and show ‘Unknown’ explicitly in reports.
4. Dates in the wrong format or time zone
What happens: 04/05/2026 is read as 5 April instead of 4 May, or timestamps stored in UTC push late-evening South African sales (UTC+2) into the wrong day, which at month end means the wrong month.
Detect: check the minimum and maximum dates, and look at sales by hour of day for odd patterns. Prevent: store ISO dates (YYYY-MM-DD) and convert time zones deliberately.
5. Mismatched definitions
What happens: sales says revenue is R2.4 million, finance says R2.1 million, and both are ‘right’. One includes VAT and uses order date; the other excludes VAT, deducts credit notes and uses invoice date. The meeting becomes an argument about numbers instead of decisions.
| Question to settle | Typical options |
|---|---|
| VAT | Including or excluding |
| Returns and credit notes | Deducted or not |
| Date basis | Order, invoice, payment or delivery date |
| Cancelled orders | Included or excluded |
| Active customer | Ordered in last 30, 90 or 365 days |
Prevent: write the definitions down once, put them in the report footer, and use the same definitions in every report. See how to choose KPIs for defining measures well.
6. Stale or partial data
What happens: the report runs before last night's sync finished, so yesterday looks like a disaster. Or one branch's data did not upload and total sales appear to drop 20%.
Detect: show the latest date in the data on every report, and compare row counts per source with the usual volume. Prevent: schedule reports after data refreshes complete, and alert when a source is missing.
7. Wrong joins between tables
What happens: joining orders to order lines and then summing the order total repeats that total once per line, so a three-line order counts three times. Or an inner join silently drops orders whose customer is missing from the customer table.
Detect: compare row counts and totals before and after each join. Prevent: sum at the right level (sum line amounts, not order totals, after joining to lines) and use left joins when you must keep every record.
8. Outliers and test data
What happens: a R999,999 test transaction, a staff account used for training, or a price typo shifts the average and the chart scale. See how to spot anomalies in business data for methods.
Detect: review the top and bottom 10 values in each amount column. Prevent: flag test and internal accounts in the source system and exclude them by rule.
Every Summarix report includes data-quality notes, so blanks, duplicates and suspicious values are visible before anyone acts on the numbers.
Free plan: 5 AI reports a month, no card needed.
Why these problems stay hidden
Each of these issues produces a number that is off by a few percent, not by an order of magnitude. A report showing R2.46 million instead of R2.31 million does not look broken, so nobody questions it. Over months, decisions get made on the wrong figures, and when the discrepancy is eventually noticed, trust in every report suffers. The fix is not heroic effort; it is a short, routine set of checks run every time, ideally built into the reporting process so nobody has to remember.
A five-minute pre-send check
- Does the headline total reconcile to a trusted source?
- Do category subtotals add up to the grand total?
- Is the latest date in the data what you expect?
- Are there duplicate keys or unexpected blanks?
- Are definitions (VAT, date basis, returns) stated?
Build this into your monthly business report routine and most quiet errors get caught before they reach a decision-maker. For the fixes themselves, work through the data cleaning checklist. Summarix reports include data-quality notes alongside the KPIs and charts; see features.
Frequently asked questions
What are the most common data quality issues?
Duplicates, inconsistent categories, missing values, date and time-zone errors, mismatched definitions, stale data, bad joins and outliers or test records.
How do data quality issues affect reports?
They produce figures that look plausible but are wrong, such as inflated totals or missing segments, which can lead to poor decisions and lost trust in reporting.
How do I check data quality quickly?
Reconcile the main total to a trusted source, check subtotals add up, count duplicates and blanks, and confirm the date range is complete.
Why do sales and finance report different revenue?
Usually different definitions: VAT included or not, credit notes deducted or not, and order date versus invoice date. Agree and document one definition.