How we measured Black Friday and Cyber Monday 2026
Published November 20, 2026 on greatgeardeals. Methodology version 1.0, fixed on November 13, 2026, and published on November 20, 2026, before the sales began.
The question we are testing
Did Black Friday and Cyber Monday offer meaningfully better buying opportunities for the products we tracked, compared with buying earlier in November? We will look at which kinds of products and which retailers benefited, whether Cyber Monday improved on Black Friday, and where waiting brought little extra discount or left fewer products in stock. These are questions, not conclusions we have decided in advance. The answer may be that buying earlier was just as good. Everything we say is about the products we tracked, not about every product these retailers sell.
The short version
- We checked the advertised prices of 2,007 gear products at 5 Canadian and US retailers, once or several times a day, from October 27 to December 2, 2026.
- We compare each price we saw during Black Friday to Cyber Monday with the prices we had seen before the event, which we call our observed baseline, and with the best price we saw between November 20 and November 26. We do not compare with the "was" or "regular" price a retailer shows, though we show that too.
- For each product we say two separate things: how far below our baseline the price went, and whether it reached the lowest price we have tracked for that product.
- We cannot see what people paid, what was in stock for them, or some conditions that apply to an offer. "What we can't see" lists them.
What we tracked
We tracked gear only: skis and snowboards, cycling, backpacks and bags, climbing gear, tents and shelters, camp and cook, sleeping gear, avalanche safety. We did not track clothing or footwear. Ski and snowboard boots count as gear. Alpine touring gets its own line when we have enough products to say anything about it.
We followed four groups of products at the same retailers, and we report them separately:
- Products that were on sale when we started. We froze this list on November 13, 2026: 780 products.
- A comparison group. Before the event we drew 226 products at random from products we already knew about at the same retailers that were not on sale, plus 65 spares, and fixed in advance how we drew them (random seed
bf2026-panel-v1). We kept checking all of them. On November 20 we counted as the comparison group only the products that met these conditions on at least 14 of the 21 days before the event: available, and priced at least 85% of the retailer's own reference price (70 of 226 places filled; where a drawn product did not meet the conditions, a spare from the same retailer and category that did took its place). Where the page showed no reference price we cannot tell whether the price was already marked down; we still count those days as regular-price days, and we report those products as a separate part of the group so the limit is visible. Many of the products we drew were sold out or already marked down, and are not in the group. This is not a representative sample of ordinary prices: it is previously known products that met the conditions, so it leaves out what had sold through and probably over-represents slower-moving products. Fewer than 10 products qualified at each of Steep & Cheap and The Last Hunt, so we show no comparison group figures for them. - Markdowns our sites did not feature. Some of the products we drew were marked down by more than 15% against the retailer's own reference price on most days before the event, although they were not on our sites. We report them as their own line (23 products) and never mix them into the other groups.
- A sample of the retailers' catalogues. Because we had only been watching products that were on sale at some point, we also drew 1,100 products at all prices from the retailers' gear category pages. We publish the whole frame on this site, at /black-friday-2026/frame/: the page addresses, each product's place in its list, what we excluded and why, and which products were drawn or kept as reserves, with the random seed (
bf2026-catalog-v1). We collected the pages on November 1, 2026. We never dropped any of them: each one was due for a check every day through December 2, including those whose price did not move, went up, or sold out. Because we sampled a different share of each retailer's catalogue (The Last Hunt 100%, Evo 23%, MEC 97%, Backcountry 12%), we report each retailer separately; any combined figure is weighted by those shares and labelled "weighted". Steep & Cheap is not in the catalogue sample, because it is an outlet with no ordinary price to measure against. Its deal-watchlist products are shown in an outlet line, measured against their own observed history.
Retailers: The Last Hunt, Evo, MEC, Steep & Cheap, Backcountry. We chose the top five by the number of gear products on sale with us on October 25, 2026. That is a statement about our own coverage, not about market share.
What we looked at
For each product we recorded the public price advertised on the retailer's own product page, in the currency the page showed (Canadian or US dollars; we never add the two together, and percentages are the same in either), along with whether the product was available to buy, the variant and size shown, and the "was" or "regular" price if the page showed one. We checked from the United States for Evo, and from a location our checking service chose for The Last Hunt, MEC, Steep & Cheap and Backcountry (we did not set one; each check records the currency the page showed). We recorded the time of every check, and we assign every check to a calendar day in Eastern time from its own timestamp, and we count distinct Eastern days, never runs.
Some prices only apply with a coupon code, a membership, a minimum spend, or a particular delivery option. We record what we can detect and report those separately. A conditional price is never mixed into the public-price figures.
What we keep as evidence. From October 28 we keep the full response we received for every check, so any number in this report can be re-read from what we stored without asking the retailer again. Checks from October 26 to October 28 were kept only as a fingerprint of the response and a short record of the fields we read. For some products we also keep a page screenshot as evidence.
When we looked
| Period | Dates (Eastern time) | How often |
|---|---|---|
| Before the event | October 27 to November 19 | once a day, every product |
| Run-up | November 20 to November 26 | every 6 hours, every product (our schedule never checks less often than this during the event) |
| Black Friday to Cyber Monday | November 26 to December 1 | every 6 hours, and every hour from 7 pm on November 25 to 6 pm on December 1 for 51 of the products that were on sale when we started, drawn at random within each retailer and category |
| After | December 2 | every 6 hours, with the last check starting at 3:20 pm |
The catalogue sample is the exception: we checked it once a day throughout, and a second time, in a different hour, on Black Friday and Cyber Monday.
These are the times we scheduled checks. When a retailer refused several of our checks in a row, we stopped checking it for the rest of that run and tried again at the next one, so on some days some products there were not checked; those days have no price and are left out, not filled in.
A deal we found already running at our first check in the run-up is labelled "started before our window": we cannot say when it began.
Our observed baseline
For each product we use the price we observed advertised, once per Eastern day: if we checked several times in a day, we take the middle value of that day's valid checks. Each day counts equally. Days we did not see a valid price are left out, not filled in.
The baseline is the middle value of those daily prices over the 21 days ending November 19, 2026. We need at least 14 days with a valid price in that window. A product with fewer is listed as having no baseline and is left out of the savings numbers (it still appears in the tables). We also show the product's October history, the last valid price before November 20, and its best price between November 20 and November 26 next to the baseline, so a long early promotion is visible and not hidden inside a middle value.
We call this "our observed baseline" on purpose. It is the price we saw advertised, not what anyone paid, and it is not a test of whether any retailer's advertised "regular" price meets any legal standard.
Two comparisons, set before we saw the data
For each product we take its lowest price during Black Friday to Cyber Monday (Friday, November 27 to Monday, November 30) that we saw in at least two separate checks made in different hours. A price we saw once is unconfirmed: we show it but do not count it in the headline numbers.
- Against our baseline. How far below the baseline was that price? The actual reduction is shown, not only a label.
- Against the best price before the event. The best price we saw from November 20 at midnight to November 26 at 11:59 pm Eastern. This answers: was Black Friday really better than buying earlier in November? We also report the share of products whose best price from November 20 to November 26 was at or below, or within 1% of, their Black Friday to Cyber Monday price, out of the products with a confirmed price in that window, because for those products waiting for the event brought nothing. A product with no confirmed price from November 20 to November 26 is counted as not having been there, so this figure is a floor.
- Friday against Monday. For products that were available on both days, we compare each product's lowest confirmed price on Friday with its lowest on Monday, using the same checks for every retailer. Prices within 1% count as the same. We also report how many products were available on Friday but not on Monday, and the reverse, so a product that sold out does not quietly disappear from the comparison. Days we have no price for stay missing: a product whose checks were refused or failed on one of the days is counted as having no reading that day, not as sold out.
Did waiting help?
For each product we also look at its lowest confirmed price on a single day, on November 20, on Black Friday, on Cyber Monday and on December 2, and compare it with our baseline: cheaper if it was more than 1% below, the same within 1%, higher if more than 1% above, and no confirmed price if it was sold out, unavailable, seen in only one check that day, or our checks that day were refused or returned no price. This uses a 1% line, not the 5% line of "meaningfully lower", so it says "cheaper than our baseline" and nothing more. We also report the share of products that were cheaper than on Friday by more than 1% on at least one of Monday, Tuesday or Wednesday, shown next to the products that were available on Friday but not later, so a product that sold out does not count as having got no cheaper. Only products with a baseline are included, and we show how many.
Two labels for every product
We give each product two labels that do not depend on each other.
1. Versus our baseline
| Label | What it means |
|---|---|
| Meaningfully lower | 5% or more below the baseline |
| Slightly lower | lower, but by less than 5% (more than 1% below) |
| Unchanged | within 1% of the baseline |
| Higher | more than 1% above the baseline |
Together, "slightly lower", "unchanged" and "higher" are "less than 5% below baseline". The 5% line is a line we drew before the data: it is not proof that smaller changes are noise, which is why we show the actual reductions too.
2. Tracked-low status
What we call a tracked low is the lowest price we saw in at least two checks since we started watching the product, October 27, 2026; it says nothing about earlier years.
| Label | What it means |
|---|---|
| New tracked low | more than 1% below the lowest price we had tracked for it |
| Matched tracked low | within 1% of the lowest price we had tracked |
| Above its tracked low | cheaper than that before |
| Insufficient history | we had watched it on fewer than 30 separate days before November 20, so we do not say whether it reached a tracked low |
A product can be, for example, "slightly lower" and a "new tracked low", or "unchanged" and a "matched tracked low" (it sat at its lowest price all along). The comparison group and the catalogue sample are mostly "insufficient history", because we only started watching them in the weeks before the event. The same labels, with the best price before the event as the starting point, tell you whether Black Friday beat buying earlier.
The retailer's own "was" price
Where a retailer shows a "was", "regular" or similar price, we record it and its wording. We then show two discounts side by side for the same offer: the one the retailer advertised, and the one against our baseline. Products with no reference price are shown on separate lines: we cannot tell whether their price was already marked down. We make no claim about why the two differ.
How long a deal lasted
A deal starts when we first see a confirmed price at least 5% below the baseline. It ends in one of three ways, which we report separately: the price rises back above that line; the product sells out or can no longer be bought at that price; or we lose sight of it because we can no longer get a valid price from the page. A failed check is a gap in our record, never an end. Deals still running at the end of our window are still running, not ended.
Because we check at set times, we report a deal's length as a range, not a single number. We report the share still running after 6, 24 and 72 hours. We also report "duration among completed deals we observed" (shown only when at least 30 deals completed), always next to the share still running, because completed deals alone look shorter than deals really were. Products we only started watching late are marked "not followed for the full 72 hours".
Products we found during the event
Around the event we also looked for products that were marked down but that we had not been watching. We found them in our regular reads of these retailers' sale and category pages on November 24 and December 1: the Tuesday before Black Friday and the day after Cyber Monday. We did not search for new markdowns on Black Friday or Cyber Monday themselves. For these products we have no baseline, because we had not watched them for long enough, so we cannot say whether the price was a real saving. We show them in a separate table, with the discount the retailer advertised, labelled "advertised discount only, no baseline", and mark them "new to our tracker". They are never counted in the labels above or in the headline figures. We see each of them only on the day we found it and do not follow it afterwards, so we do not report how long their deals lasted. After the event we also publish how much of what we found we had already been watching, with the limits of that count.
What we can't see
- What people paid. We see advertised prices, not sales, so we cannot say how many units sold or at what price.
- In-store prices, and prices shown only to people who are logged in, in a different region, or on an app.
- Some conditions of an offer, such as the exact terms of a coupon, membership or minimum spend, or the cost of delivery. We record what the page lets us detect and mark the rest as unknown.
- Stock levels. We see whether a product was shown as available, not how many were left.
- Anything between our checks. A price that appeared and vanished between two checks is invisible to us. Before November 20 we check once a day, and until the event no more often than every 6 hours, so a shorter sale can be missed.
- Products we did not watch. The catalogue sample is a sample; the figure above says how much of what our search found we were already watching.
- For MEC, we could see whether a product was available only on the checks where the page gave us that information; the report says where it did not.
Corrections and changes
If you think something here is wrong, write to contact@greatgeardeals.com with the product and the date. We correct errors on the page, list each correction with its date at the bottom, and never change the methodology after November 13, 2026 without saying so here. Version 1.0. Data checks run through December 2, 2026.
