Reading your Buen Fin results honestly

Buen Fin 2026 closes on Tuesday November 17. On Wednesday the 18th, the conversation inside almost every operation is the same: how much did we sell. Somebody pulls the total, compares it against last year, and if it went up there is celebration. It is the most natural reading in the world and it is the one that generates the most bad decisions for the following year.

The problem is not the enthusiasm. It is that the sales total of a promotional event is a structurally misleading number. It is inflated by three things at once: demand you were going to get anyway, demand you pulled forward from the following weeks, and a per-unit margin lower than normal. A higher total with those three factors inside can mean you earned more, or it can mean you gave away margin to move inventory that would have moved regardless.

And concrete decisions come out of that reading. How much inventory to bring next year. Which products to promote. What discount to offer. If your 2026 conclusion is “25 percent worked because we hit a record”, in 2027 you will repeat a discount you may not have needed, on inventory you over-bought, in the wrong SKUs.

This article proposes a different reading. It is not complicated and needs no statistical tooling: it needs a baseline defined before you look at the result, and the willingness to accept whatever comes out. That is the difference between a report that feels good and one that is useful next year.

why the event total always impresses

A promotional event concentrates into five days demand that normally spreads across several weeks. That alone, without you doing anything different, produces a total that looks spectacular against any normal period of the year.

On top of that sits the discount’s effect on units: lowering price raises quantity sold, almost always. And on top of that sits the extra advertising most of us pour into those days. All three push the same number upward, and none of the three tells you whether you made more money.

That is why gross sales is, of all the figures you will look at on November 18, the least informative one. It is the same underlying problem covered in gross sales vs net sales, with high season acting as an amplifier.

separating incremental sales from the ones you had anyway

The honest question is how much of that volume would not have happened without the event. You answer it with a baseline and a subtraction, like this.

An example with numbers. Over the five days you placed 1,400 units of one product family. Across the four previous weeks, that same family averaged 95 units a day. Five days times 95 is 475 baseline units: the ones you would presumably have sold with no promotion. Gross incremental sales are 1,400 minus 475, so 925 units.

So far it sounds great: you nearly tripled. But margin is not the same. Suppose that under normal conditions that product leaves you 240 pesos of net margin per unit, and with the event discount it left you 110.

Now the two calculations. Event profit is 1,400 units times 110 pesos: 154,000. The profit you would have had without the event is 475 units times 240 pesos: 114,000. Real incremental gain is 40,000 pesos, not the 154,000 in the headline.

Seen another way, which teaches more: the 925 incremental units contributed 925 times 110, or 101,750 pesos. But the 475 baseline units, which you would have sold at 240 of margin, you sold at 110, so you gave up 130 pesos on each one: 475 times 130 is 61,750 pesos of margin handed over. 101,750 minus 61,750 is the same 40,000. The discount paid for itself, but it left far less than the total suggested.

Glossary: real net margin, with everything deducted →

how to set a baseline that does not fool you

The baseline is the most debatable part of the exercise and it is worth being explicit about its limits. This is not a controlled experiment: you have no control group and you cannot know what would have happened. It is an estimate, and what makes it useful is defining it before you see the result and using it the same way every year.

Three criteria help. First, use a prior period long enough to smooth out noise, four weeks is usually enough, and strip out any day when you were already running a promotion or when you had a stockout, because a sold-out day understates your normal demand. Second, do it per product family or per SKU, not for the whole business: mix shifts so much during the event that a global average means nothing, as explained in analyzing sales by product family and season. Third, if you have prior years, use them to correct for trend: if you have been growing, your November baseline should be higher than your October average.

And write it down. The value here compounds: the 2026 analysis matters mostly because it gives you the 2027 comparison point. That is the argument of the value of your history applied to seasonality.

the hangover effect, which almost nobody subtracts

One adjustment is missing and it is the one that changes the result the most. Part of the demand you served during the event was not new demand: it was demand from the following weeks that the customer pulled forward because there was a discount. You did not win that sale, you moved its date, and you moved it to a lower margin on top.

You measure it by looking at the weeks after. With the same example: if your baseline was 95 units a day and the week following the event averaged 62, you were short 33 a day for seven days, or 231 units. Those 231 are demand that was pulled forward.

Redo the math with the adjustment. Of the 925 incremental units, 231 were not new, so the honest increment is 694 units. And those 231 were also sold at 110 of margin instead of 240, so there is another 231 times 130, or 30,030 pesos, of margin forgone. Incremental gain falls from 40,000 to 9,970 pesos.

Just under ten thousand pesos of incremental profit against 154,000 of reported profit. That does not mean the event was a bad idea: it means its real benefit was a quarter of the apparent one, and that next year it is worth checking whether the discount had to be that deep or whether half of it plus visibility would have done. That is exactly the decision a sales total hides from you, and the mechanism behind selling more but earning less.

Two cautions about this adjustment. The post-event drop can have other causes, such as a stockout of your own or a competitor launching a promotion afterwards, so check before attributing all of it to pull-forward. And the observation window has to stay consistent: if you use one week in 2026, use one week in 2027.

why the best-seller ranking misleads

The other report everybody pulls on November 18 is the Buen Fin best-seller ranking. It is useful for understanding traffic and terrible for deciding inventory, because it is sorted by units and units are not money.

An example. Your top-selling SKU of the event placed 620 units at a net margin of 38 pesos each: 23,560 pesos of profit. Another SKU, sitting far down the ranking, placed 95 units at 520 pesos of margin: 49,400 pesos. The second contributed more than twice the profit while selling a sixth of the units.

If your 2027 purchasing decision comes out of that ranking, you will bring more of the first and the same of the second, which is precisely the allocation that leaves money on the table. And if the first one also consumed the bulk of your ad budget, the ranking is charging you twice.

The fix is to sort the same report by total net profit contributed and, in a second column, by net profit per unit. The first order tells you what carried the event; the second tells you what deserves more inventory and more advertising next year. When the two orders agree, you have a genuine winner. When they disagree, that is your decision.

Glossary: sales velocity, the pace at which each SKU moves →

the five questions actually worth answering

So the analysis ends in decisions rather than in a slide deck, reduce it to questions with actionable answers.

How many units were genuinely incremental, after subtracting baseline and pull-forward. How much incremental net profit the event left, compared against the no-event scenario. Which SKUs contributed profit and which contributed units, with both ranking orders side by side. Which SKUs ran out before the close and how much demand went unserved, because that is the only cost that appears in no report at all. And what discount each SKU actually needed: if one product sold well at 15 percent and another needed 30, they do not have to carry the same one next year.

There is a sixth that is not answered on November 18 but in December: how much inventory is left, how much capital is parked in it and how fast it is moving now that the peak is over. That is the question that decides whether December goes well or whether you start January with a full warehouse.

what this analysis looks like in iqseller

In iqseller the tedious part of this exercise, which is pulling both channels’ numbers under the same definition, is already done. The Analytics module shows sales, units and orders per SKU and per channel with a prior-period comparison, which is the baseline from the example, and the report builder lets you assemble the per-family view without exporting to a spreadsheet, as covered in report builder.

The Profitability module is what makes the ranking honest: it computes net margin per SKU with the COGS you load, commissions from the Amazon settlement, commissions on MercadoLibre orders, FBA and Full fees, shipping, advertising and the VAT and withholding breakdown. With that, sorting by profit contributed rather than by units is a matter of switching columns. The Parent, Model and SKU tree prevents the classic mistake of analyzing at the parent level when one variant explains the whole result.

For the December question, Inventory shows the valuation of the stock left over and Forecast shows coverage against current velocity, which after the peak is usually very different from the event’s.

What iqseller does not do is compute incremental sales for you. There is no baseline button and it does not estimate what you would have sold without a promotion, because that rests on an assumption you make. What it gives you are the two figures the subtraction needs, under the same definition in both channels, without you assembling them by hand.

the half-page document that actually helps

Close the analysis in half a page and store it somewhere you will find again in November 2027. Which baseline you used and why. How many net incremental units came out. How much incremental profit. The three SKUs that contributed the most profit and the three that contributed the most units, flagging where they disagree. The SKUs that ran out, and when. And the discount that was genuinely needed in each case.

None of that is a pretty report, and that is why it gets used. Next year you will not need to remember how much you sold. You will need to know how much of it was worth it. That half page, written while the numbers are fresh, is what turns five days of effort into a better decision the next time around.