Why the budget question comes after Black Week
In 2026, Black Friday falls on November 27 and Cyber Monday on November 30. Both sale days are therefore in the same calendar month, unlike 2025, when Cyber Monday fell on December 1. Demand is high: around one in two (52 percent) of the surveyed internet users aged 16 and over wanted to use the reduced prices of Black Friday 2025, the vast majority of them online (Bitkom 2025). Two thirds (69 percent) intended to look for deals exclusively online, only 4 percent exclusively in local stores (Bitkom 2025).
Many come with a fixed amount: the majority (61 percent, compared with 53 percent in 2024) already had a clear idea of how much they wanted to spend, 312 euros on average and thus around 11 percent more than the 280 euros of 2024 (Bitkom 2025). For retailers this means ad budgets rise in these weeks, and with them the question of which campaign earned its money back. Neither gut feeling nor revenue alone answers that question. A campaign with high revenue may have cost more than a quiet one with small revenue and almost no costs. Only the ratio shows where the next euro belongs.
A ROAS of 4 means: every euro of ad spend brought four euros of revenue, as captured by the tracker. Cost of goods, shipping, payment fees and returns are not included. Whether a campaign makes money only shows in comparison with your break-even ROAS, which this article derives further below.
How XICTRAQ® calculates ROAS per campaign
In XICTRAQ®, ROAS sits in the analytics area under e-commerce at “Ad-Spend & ROAS”, and there is also a “ROAS by campaign” card for the freely configurable overview. The calculation brings together two data sources that come about separately: the ad costs you enter yourself per campaign and month, and the revenue that cookieless e-commerce tracking measures in the sessions of your shop. Which orders count and how sessions come about is described in the analytics documentation (XICTRAQ documentation). Shops without automatic detection report their purchases via the JavaScript tracker. Like the other revenue reports, the revenue module with ROAS is included from the Business plan upward.
The calculation has four steps, and each of them shapes what ends up in the table:
- Revenue per session: All purchases in the selected period are grouped by session.
- Campaign of the session: Each session gets the first non-empty
utm_campaignthat occurred in it. If no page of the session carries a campaign name, its revenue counts under “(direct)”. - Costs in the period: From each monthly row, the share of the month covered by the selected period counts.
- ROAS: Revenue of the campaign divided by its pro rata costs. Without a cost entry the field stays empty instead of faking an infinitely good campaign.
Two consequences of these rules matter for Black Week. First, “(direct)” in this table is not a direct visit in the narrow sense but the catch-all row for every purchase from a session without a campaign name. Purchases after an organic search, after a link in a forum, after a click in a newsletter without campaign parameters and after a visit via a bookmark end up there. Second, the first campaign in a session counts: someone who arrives via the newsletter in the morning and returns via an ad in the afternoon credits the purchase to the newsletter, as long as both are in the same session.
| Column | Where the value comes from | What to watch |
|---|---|---|
| Spend | Your entries per campaign and month, allocated pro rata to the period | Without an entry the column shows 0 and ROAS stays empty: not recorded does not mean free |
| Revenue | Purchases from sessions whose first campaign is this one | Order value according to the tracker, without deducting cost of goods |
| Conversions | Sessions with at least one purchase | Two orders in one session make one conversion |
| ROAS | Revenue divided by pro rata spend | Empty if no cost entry for the campaign falls in the period |
| (direct) | Purchases from sessions without a campaign name | Also includes search, referrals and newsletters without parameters |
| Previous period | Equally long time window directly before | Missing if revenue was only recorded later |
Attribution per session: where measurement ends
XICTRAQ® works without cookies and without an identifier in the browser. Sessions are merged via a pseudonymous daily identifier, a hash of the IP address and a random value generated anew each day and deleted once the day is over (XICTRAQ documentation). For reach measurement, this is why it usually works without consent: neither a cookie nor an identifier is stored in the browser, and the law ties the consent requirement to storing information on the device and accessing it (TDDDG, Section 25 (1)). The assessment of the individual case remains with you.
For ROAS, the daily identifier has a clear consequence that follows from the method itself: a session ends at the latest when the day changes, and it depends on the address through which a device connects to the internet. If someone clicks an ad on Monday and buys on Friday, those are two sessions. If the second one carries no campaign name, the purchase appears under “(direct)” and not with the campaign that triggered it. The same applies when someone clicks the ad on the go via the mobile network and orders in the evening via the home connection: new address, new session. Conversely, devices that share an address, for example in an office, flow into the same session.
No study measures how often the click and the purchase are separated in your shop. The Bitkom survey does show, however, that by their own assessment many people plan their sale purchases: 45 percent of those who wanted to shop on Black Friday or during Cyber Week saw themselves as “wish fulfillers” (Bitkom’s category, translated), who treat themselves to things they have wanted for some time but do not want to buy at full price (Bitkom 2025). Just as many (45 percent) counted themselves among the “deliberate buyers”, who specifically buy everyday items or make larger purchases when these are on sale (Bitkom 2025). Only 28 percent saw themselves as “impulse buyers” (Bitkom 2025), and 33 percent used the discounts mainly for gifts for family and friends (Bitkom 2025). Multiple answers were possible.
The following conclusion is ours, not Bitkom’s: someone who waits for a specific price often sees the ad days before the purchase. Attribution per session spreads exactly these purchases across several rows. Device switching comes on top. For online shopping, 73 percent use a smartphone, 52 percent a notebook, 31 percent a desktop computer and 27 percent a tablet (Bitkom 2025). Together the shares add up to far more than 100 percent, so many people use more than one device. Whether a device switch also means a new session depends on the network: behind a shared home connection, devices often share one address, while on the mobile network the same phone has a different one than at home.
Anyone who stumbles upon a bargain unplanned would do well to compare the price once more – online, an overview is only a few clicks away.
— Dr. Rohleder, Managing Director of Bitkom, press release of November 24, 2025, translated from German
For measurement, a price comparison is another break. Someone who leaves the shop to compare and later returns via a search engine or a bookmark no longer brings a campaign name along. On the same day and via the same address, the first campaign of the session still counts; on another day it does not.
Attribution per session is strict with campaigns that act early in the decision path and generous with those clicked shortly before the purchase. A search ad from Monday may be undervalued, a reminder newsletter on sale day overvalued. Read the “(direct)” row as part of the result: if it rises much more during Black Week than in the weeks before, it probably contains revenue that your campaigns triggered. Visits from AI assistants also usually arrive without a campaign name; how to measure this channel separately is shown in the article on referral traffic from AI recommendations.
Monthly costs: the trap with the sale period
In XICTRAQ®, you enter ad costs per campaign and calendar month, not per day. When you select a period in the dashboard, the share of each month covered by the period counts. The interface says so itself: monthly costs count pro rata by days in the range. That is correct as long as costs accrue evenly across the month, and it prevents another error: seven days of revenue against the costs of an entire month would otherwise push ROAS down to a fraction. During Black Week, however, costs do not accrue evenly.
Suppose your shop schedules its Black Week for November 20 to 30, and the search campaign runs only on these eleven days. If you select exactly these days, XICTRAQ® counts eleven of thirty November days, i.e. 11/30 of the November costs entered. The rest of the budget is left out, even though it was spent in the same week.
Campaign bf-search: November costs of 3,000 euros, spent entirely between November 20 and 30, measured revenue of 8,250 euros. If you select only November 20 to 30, 11/30 of the costs count, i.e. 1,100 euros, and ROAS is 7.5. Over all of November, the full 3,000 euros count and ROAS is 2.75. Revenue is the same in both cases, ROAS differs by a factor of 30/11, almost threefold. The correct figure here is the monthly value.
This leads to three rules for the evaluation:
- Separate campaign names for sale week. If Black Week ads run under the name of the ongoing campaign, budget and revenue of both phases mix in one monthly row. With separate names such as
bf-searchorbf-newsletter, the sale budget stands on its own. - Evaluate the whole month. If a campaign runs only in November, the period from November 1 to 30 covers its costs completely, and outside its run time hardly any visits arrive with its name.
- Split costs across the month boundary. If a campaign extends into December, enter two rows, each with the costs actually incurred in that month. In 2026, Black Friday and Cyber Monday are in November; in 2025, Cyber Monday fell on December 1. What this calendar shift means for reports is described in the article on year-over-year comparisons in the fourth quarter.
Campaign names: link and cost list must match
Costs find their revenue only via the campaign name. If the link says utm_campaign=BF-Search and the cost list says bf-search, the table shows two rows: one with costs and no revenue, one with revenue and no costs, and neither has a usable ROAS. Upper and lower case, hyphens and spaces must therefore match exactly. What a naming convention that holds up in daily work looks like, and where parameters get lost along the way, is shown in the article on clean tagging with UTM parameters.
A second reason for clean parameters lies in the browser. Without a policy set by the website, the referrer policy strict-origin-when-cross-origin applies (MDN). When navigating to another website, the browser then sends at most the origin, i.e. protocol and domain, not path and parameters (MDN). A campaign therefore cannot be recognized from the referrer alone, and clicks from email programs and apps often arrive without any referrer at all. The campaign name has to be in the target link itself, for example like this:
https://shop.example/sale?utm_source=newsletter&utm_medium=email&utm_campaign=bf-newsletter
https://shop.example/sale?utm_source=search&utm_medium=cpc&utm_campaign=bf-search
https://shop.example/sale?utm_source=social&utm_medium=paid-social&utm_campaign=bf-socialXICTRAQ® also captures click identifiers such as gclid or fbclid that ad networks append to links. The ROAS table, however, groups by utm_campaign. Without campaign parameters, an ad click therefore lands in the “(direct)” row, even if a click identifier came along.
Recording costs: form or CSV
You enter costs under e-commerce at “Ad-Spend & ROAS”: campaign, optionally the source, cost and month. For several campaigns, the CSV import is faster. One row per campaign and month is enough, amounts with a period as decimal separator, the month in year-month format:
campaign,source,cost,month
bf-newsletter,newsletter,400,2026-11
bf-search,search,3000,2026-11
bf-social,social,1200,2026-11Important for Black Week: campaign and month together form the key. A second entry with the same campaign and the same month replaces the amount, it does not add to it. If you enter an interim total halfway through, simply overwrite it after month end with the final amount from the invoice. Anyone who imports daily costs one by one instead keeps only the last day entered.
Decide once what belongs in the cost amount and stick to it: only the media costs of the ad networks, or also design and agency fees. For comparing campaigns, what matters most is that all are recorded by the same rule. Discounts that already lower the order value do not belong in the costs again, otherwise they count twice.
“Ad-Spend & ROAS” calculates without filters by country, device or channel, because costs exist only per campaign and month. If revenue were narrowed to one device while costs stayed whole, ROAS would drop systematically. Date limits, on the other hand, apply to costs and revenue alike.
Reading ROAS: break-even instead of gut feeling
Whether a ROAS of 2.75 is good depends on your margin. The break-even ROAS is the value from which a campaign’s gross profit covers its ad costs. It is 1 divided by the share of revenue left after cost of goods, shipping and payment fees. If 30 euros remain from 100 euros of revenue, break-even is 1 / 0.30, around 3.3. A campaign with a ROAS of 2.75 has cost money in this case, even though it brought 2.75 euros of revenue for every euro of advertising.
| Share of revenue left over | Break-even ROAS | ROAS 2.75 is then |
|---|---|---|
| 20 % | 5.0 | well below |
| 25 % | 4.0 | below |
| 30 % | 3.3 | below |
| 40 % | 2.5 | just above |
| 50 % | 2.0 | above |
The “ROAS by campaign” card colors campaigns green from a threshold you set in the card settings; without your own value it is 1. If you enter your break-even value there, the color shows what pays off and not just what had more revenue than costs.
Sale week shifts two more things. Discounts lower the margin, so the break-even ROAS rises precisely when the campaigns run. And ROAS works with the order value at the time of purchase; it does not show how much remains after returns. For that, it is worth looking at the return rate of the sale items in the shop. Orders that XICTRAQ® imports from Shopware 6 via the shop connection and that have no underlying visit count in revenue but stay out of ROAS (XICTRAQ documentation). How to reconcile the shop and analytics is covered in the article on reconciliation via order import.
Previous period and previous year: what the comparison shows
Alongside the selected period, the report delivers the revenue of the equally long previous period directly before it. For November, that is essentially October. With separate Black Week names, the previous period usually has no revenue per campaign, so the comparison says little there. It is meaningful for the total and for ongoing campaigns that ran in both periods. If revenue measurement for your website only began after the start of the previous period, XICTRAQ® shows no comparison instead of a false zero.
The really interesting comparison is the one with the previous year’s Black Week. It requires that last year’s data is still there: in the Business plan, the first plan that includes the revenue module, analytics history is unlimited according to the pricing page (XICTRAQ pricing page). Because Black Friday falls on a different calendar day every year, it is worth comparing by sale day rather than by date. Keep campaign names stable over the years for this, for example bf-search in 2026 and 2027, and note the cost rule you calculated with.
Schedule for Black Week 2026
The following schedule keeps measurement clean without creating work during sale week itself. The dates depend on your shop’s sale period; the example assumes a start on November 20.
- Up to one week before the start: Define campaign names, for example
bf-plus channel, addutm_campaignto all target links and share the list with everyone who creates ads or newsletters. - Before the start: Open each link once and check under sources and campaigns whether the name arrives. Whether the shop reports purchases cleanly is checked as described in the article on conversion tracking without third-party scripts.
- During sale week: Look at revenue and sessions per campaign daily, but do not judge ROAS yet. Monthly costs are only complete at the end.
- After Cyber Monday: Enter the actual November costs from the invoices and overwrite preliminary values.
- Early December: Evaluate the period November 1 to 30, hold ROAS against break-even and compare the “(direct)” row with the weeks before.
- For next year: Note campaign names, cost rule and result so that Black Week 2027 is measured by the same rules.
Anyone who wants to keep an eye on competitor prices in the same week will find a separate article on tracking pricing pages for Black Week. For the question of return per campaign, the table above is enough, provided names and costs are right.
Making return per campaign visible with XICTRAQ®
Black Week ROAS can be calculated reliably without cookies if three things fit together: campaign names spelled the same in the link and in the cost list, costs per month with a reporting period covering the whole month, and a look at the “(direct)” row, which catches what attribution per session cannot connect. XICTRAQ® brings costs and measured revenue together per campaign and shows which campaign earned back its budget. The live demo shows what the report looks like. The basics of tracking are described in the article on e-commerce tracking without cookies, and if you are starting out, you can register directly.
This article is based on: Bitkom, press release “Black Friday: Die Hälfte geht auf Schnäppchenjagd” of November 24, 2025, representative survey by Bitkom Research of 1,120 internet users aged 16 and over in Germany, including 1,072 online shoppers, survey period calendar weeks 40 to 41 of 2025; Telecommunications Digital Services Data Protection Act (TDDDG), Section 25 (1); MDN Web Docs, page on the Referrer-Policy header; XICTRAQ documentation, analytics; XICTRAQ pricing page. The worked examples are based on assumed values. Attribution and cost allocation are described based on the XICTRAQ® method; the article does not replace business or legal advice.