Why Did Sales Drop? A 6-Step Data Diagnosis

Monday, eight in the morning. A manager opens the monthly report and sees a shorter bar than last month. Within fifteen minutes the conversation has already covered "the sales team lost its edge", "the market is frozen" and "maybe it's time to change agencies". None of those sentences came from a number.

The pattern repeats at companies of every size. The drop shows up in a single line, revenue, and the discussion starts with the causes that are easiest to blame rather than the ones most likely to be true. Three weeks later someone discovers that cost per click rose 50% while the budget had been frozen since January. The sales team had nothing to do with it.

Finding out why sales dropped is a job of elimination, not intuition. And it can be done in a single morning with data the company already has.

Why did sales drop? The short answer

Sales dropped for one of four reasons: fewer people arriving, fewer of those people buying, each purchase being worth less, or a shift in the product and channel mix. To find out which, break revenue into visitors × conversion rate × average order value and compare each factor against the same period last year.

That decomposition is the only honest shortcut. It does not name the cause, but it eliminates three quarters of the hypotheses in ten minutes and tells you where the rest of the morning is worth spending.

The arithmetic behind the diagnosis

Revenue is always a product of factors, never a standalone number. In retail and e-commerce: visitors × conversion rate × average order value. In B2B sales: leads × close rate × average contract value. In subscription businesses: customer base × revenue per customer, with churn pulling the base down.

The usefulness here is arithmetic. If revenue fell 15% and you can measure the three factors, the drop has to show up in at least one of them. If it shows up in none, you have a measurement problem, not a sales problem.

Step 1: confirm the drop is real

A good share of "drops" are noise or a badly chosen comparison base. Two questions settle it: is the decline larger than the normal variation of the last 12 months? And was the previous month typical, or was it a peak?

Brazilian retail offers a clean illustration. In April 2026, retail sales volume fell 1.5% against March in the seasonally adjusted series, according to the Monthly Retail Survey from IBGE, Brazil's national statistics office. In that same month, compared with April 2025, the sector grew 1.0%. Same data, opposite signals, purely because the base changed.

In practice: always run both comparisons. Month over month captures recent movement; year over year neutralises seasonality. When both point down, the drop is real. When they disagree, the problem is usually your reference month, not your company.

Step 2: break revenue into its parts

Take the month of the drop and the comparison month, then calculate the three factors. Here is a worked example from a mid-sized home and décor online store:

Factor	August	September	Change
Visitors	48,000	41,000	-14.6%
Conversion rate	1.80%	1.70%	-5.6%
Average order value	$320	$335	+4.7%
Revenue	$276,480	$233,495	-15.5%

The reading is immediate. Order value actually went up. Conversion barely moved. Almost the entire 15.5% drop came from traffic. Any debate about pricing, promotions or sales scripts is over before it starts.

This is where most meetings go wrong: someone sees conversion sliding from 1.80% to 1.70% and proposes a landing page optimisation project. Sensible in isolation, but it would address 5.6% of a 15.5% problem. If you want to go deeper on each factor, it is worth reviewing how to calculate and improve conversion rate and what actually moves average order value.

Step 3: drill one level into the guilty factor

Traffic fell. Traffic from where? Splitting by channel usually closes the case:

Channel	August	September	Change
Paid media	19,200	12,800	-33.3%
Organic search	14,400	14,100	-2.1%
Direct	8,400	8,300	-1.2%
Email and CRM	6,000	5,800	-3.3%

Of the 7,000 visitors lost, 6,400 came out of paid media. More than 90% of the problem lives in one channel. Final level: the budget did not change, $38,400 in both months. What changed was cost per click, from $2.00 to $3.00. Same money, a third fewer clicks.

Notice that the diagnosis ended somewhere specific and actionable: a more expensive auction in September. The possible responses (renegotiate the budget, revise keywords, refresh creatives, accept lower volume and protect margin) are business decisions. But now they are decisions about the right problem.

Step 4: separate the market from yourself

Not every drop is the operation's fault. In July 2026, Brazilian retail slipped 0.8% for the month, with furniture and appliances down 4.9%, and IBGE itself attributed part of the move to a sales hangover after the World Cup peak, as reported when the survey was released. An appliance retailer down 4% that month did not have an operational problem. It had a normal month inside a shrinking sector.

The test is simple: is your decline bigger or smaller than your sector's over the same period? If you fell 4% while the sector fell 4.9%, you gained share while shrinking. Painful for cash flow, but not a reason to rebuild the strategy. If you fell 15% while the sector held flat, the problem is internal and urgent.

In practice almost nobody runs this comparison, which is why so many companies switch agencies in a down year and switch again the year after.

Step 5: mix hides drops that totals never show

One case the method above can almost miss: revenue falls, yet every factor looks stable. The explanation is usually composition.

Brazilian e-commerce produced a sharp portrait of this in the first half of 2026. According to research by Confi/Neotrust published by E-Commerce Brasil, revenue grew 16.9% and orders rose 34.8%, while average order value fell 13.3% and items per purchase dropped 12.4%. More purchases, smaller ones. A store caught in that dynamic can watch revenue hold steady while margin collapses, because shipping and acquisition costs do not shrink at the same rate.

So beyond the three factors, look at distribution: which category lost weight in revenue? Which price band disappeared? Which customer segment pulled back? A 15% drop in total revenue can be a 60% drop in one segment nobody was watching, masked by growth somewhere else.

Step 6: rule out measurement error before acting

Before you take the diagnosis to the board, ask the uncomfortable question: did sales fall, or did measurement break? A conversion tag removed during a site update, a CRM integration that stopped syncing, a domain change that wiped history, orders that started being recorded in a different system.

The classic tell is an abrupt decline that starts on one exact day, with no transition. Real sales fall on a ramp; measurement breaks on a step. It is worth reconciling the report against the primary source, billing in the ERP or the payment gateway statement, before drawing any conclusion. Silent data quality errors account for a surprising share of crises that never existed, and a change in the attribution model is another frequent suspect when a channel "vanishes" from one month to the next.

Frequently asked questions

Sales dropped: what should I do first?

Change nothing in the first few hours. Confirm the drop using two comparison bases (the previous month and the same month last year) and break revenue into visitors, conversion and average order value. Only once you know which factor fell does any action have a chance of working.

How do I know whether a drop is seasonal or a real problem?

Compare against the same month last year, not the month immediately before. If the decline shows up in both comparisons, it is real. If it only appears against the previous month, you are probably comparing an ordinary month with a seasonal peak.

How long should I wait before reacting to a sales drop?

It depends on volume. A business with hundreds of orders a day can detect a genuine deviation in three to five days. With a handful of orders a week, one bad week is statistical noise and reacting to it costs more than waiting. The rule of thumb: wait until you have enough volume that the drop no longer fits inside the period's normal variation.

Could the drop be the market rather than my company?

It can, and it often is. Compare your change with your sector's over the same period, using public data such as national retail surveys or industry association reports. Falling less than the market means gaining share, which changes the decision entirely.

What data do I need to run this diagnosis?

The minimum is revenue, order count and visitors or leads per month, split by channel. From those you derive conversion and average order value by division. Media cost and margin per product make the diagnosis far better, but they are not a prerequisite for starting.

The diagnosis matters more than the reaction

Companies rarely fail because of a sales drop. They fail by reacting to the wrong drop: cutting the budget that was still profitable, firing the team that was not the bottleneck, discounting their way out of a traffic problem. Half an hour of decomposition before the first decision saves months of course correction.

The obstacle is usually operational rather than conceptual: the numbers live in four different places and nobody has time to rebuild the table every Monday. That friction is exactly what Sherlok removes, connecting Meta Ads, Google Ads, GA4, CRMs and spreadsheets so that "why did sales drop in September?" can be asked in plain language and answered with the decomposition already done. The analysis stays yours; what disappears is the half hour of spreadsheet work before it.