The September return is won in August, Black Friday is won in September

Tuesday 15 September 2026, 9.30 in the morning. We're in the industrial estate between Santa Maria Capua Vetere and Capua, inside Berto's office. For the first time since June, the air conditioning is off. Through the first-floor windows drifts that first bit of Campanian cool that lasts just until eleven, before the sun starts hammering the corrugated roofs of the warehouses again. There's a sharp smell of pod-machine coffee mixing with pressed cardboard and printed paper wafting up from the warehouse below.

Berto's sat at his light-formica desk. He's finally pulled out of the bottom drawer the 2026 commercial calendar I gave him back in January. It had sat there for eight months, rolled up tight, with the red elastic band that's left a dark, sticky mark on the edge of the glossy paper. He unrolls it slowly, smoothing out the corners with palms marked by thirty years of proper hard work.

"Umbè," he says, arms crossed, "I did what you said back in mid-August. Got the ideas straight, all the lads are back at work, production's running at full pelt. We're ready now. From today, we start thinking about the year-end promotions."

He's smiling. Convinced he's well ahead of schedule. In the traditional business culture Berto represents with admirable dignity, September is the month you get back into gear. The return's done, the head's clear, the route's mapped out. But as he finishes pouring the coffee into the plastic cup, I hand him my phone, screen lit up.

On the screen is a screenshot sent to me twenty minutes earlier by Bobby, straight from his studio in Manchester, a stone's throw from Salford. It's his business dashboard on Meta Business Suite. He sells gift items and leather accessories. The Black Friday and Christmas editorial calendar grid is coloured solid green, top to bottom. The product photos for the November promotions were already shot in-studio at the start of September. The ad copy's approved. The dynamic catalogues are synced and the test campaigns are already running on minimum budget to warm up the pixels.

Bobby wrapped up his Q4 strategic planning on 12 September. Over in England, they're not thinking about what to post for the feast of San Gennaro or the start of autumn — they're already simulating server load spikes for the night of 26 November.

Berto looks at the screenshot, then looks at me. Looks at his calendar unrolled on the table, trying to curl itself back up. That ten-second, awkward silence — broken only by a forklift humming away in the yard — is the exact snapshot of what happens every single year to thousands of small and medium Italian businesses.

Berto's problem isn't a lack of willingness, nor is it laziness. It's a massive error of time perception. Berto measures the time left until Black Friday by looking at the date on the calendar: 27 November. In his head, two and a half months is an eternity. In the operational reality of communications, content production, logistics — and above all, the sheer maths of how advertising algorithms work — 15 September isn't the start of preparations: it's the very last useful minute before your margin for manoeuvre shrinks down to pure damage control.

If you think Black Friday gets planned in November, you've already lost. If you think you'll get away with knocking together four graphics on Canva the week before, brace yourself to burn budget on ads nobody will see, or to deal with furious customers because the site crashes while they're trying to pay. You win the September return, yes. I wrote about that here. But Q4 — the quarter that decides whether your year closes in profit or loss — is won, irrevocably, in September.

The calendar that deceives, because November's already late

To understand why most business owners wake up too late, we need to line up the real dates for 2026 first. These aren't opinions — they're fixed deadlines around which the whole of Western commerce revolves:

• 11 November 2026 — Singles' Day (11/11). Born in China and by now firmly embedded across every global marketplace. Anyone selling on Amazon or running a B2C e-commerce site needs to treat this date as the real first taste of promotional traffic.
• 27 November 2026 — Black Friday. The absolute peak of attention and spending.
• 30 November 2026 — Cyber Monday. The digital, tech-focused tail end of the long weekend.
• 26 December 2026 — Boxing Day. A key date across the UK and the Commonwealth, kicking off a second, frightening wave of post-Christmas sales.
• Early January 2027 — the start of the winter sales in Italy. In Campania, as in most Italian regions, the regional calendar fixes the very first days of the year as the start date for the official sales.

Looking at that list, the natural human instinct is to calculate time in a straight line: "It's the 15th of September today, there's more than two months till the 27th of November, I've got plenty of time." That's the first great deception. Marketing time isn't measured from today forward to the event — it's calculated backwards, starting from the execution time each process actually needs.

Picture having to make a proper Neapolitan ragù for Sunday lunch. You can't decide to put the pot on the heat at half eleven and expect to serve, by one o'clock, a ragù that's properly slow-cooked, dark and rich. Crank the flame up to make up for lost time, and you don't get a faster ragù — you get a burnt, bitter sauce that ruins the meat and leaves everyone with an upset stomach. The only mathematically sound way to have a flawless ragù ready by one in the afternoon is to start browning the meat the evening before and let it simmer away from first light.

Communications and commercial management for an SME follow exactly the same physical law. You can't compress a campaign's development time without wrecking its quality. If you want to launch a properly structured promotion on 27 November, you have to work backwards:

1. Conversion and purchase window: 27–30 November
2. Ad warm-up phase (learning phase): 1–20 November
3. Campaign creation, upload and approval: 20–31 October
4. Production of visual content (photo shoots, video, post-production): 1–20 October
5. Defining offers, catalogue audit and profit margin: 15–30 September

Shift that first box — defining the offers — to mid-October, and the domino effect wipes out everything after it. You'll end up photographing products with your phone under the warehouse's neon strip lighting on 8 November, uploading campaigns at midnight on 24 November, and discovering on 27 November that Meta's flagged your ad account for a routine review just as your site goes offline.

The backwards calendar from Black Friday 2026

The huge mistake people who improvise make is believing these phases can overlap. They can't. Each phase is the foundation the next one stands on. If September's foundations are made of sand and rushed decisions, the whole structure collapses under the weight of November's traffic.

The technical bit almost no SME knows about: the ADS algorithm learning phase

There's a purely technical reason — down to the maths of advertising platforms — why launching campaigns right before Black Friday means handing money to Meta and Google. It's got a precise name: the learning phase.

When you create a new ad campaign on Meta (Facebook and Instagram) or Google Ads (Performance Max campaigns in particular), the algorithm doesn't know exactly which specific users will convert at the lowest cost. It has to find out. To do that, the platform automatically enters an exploration phase, showing your ads to different audience segments and analysing the responses in real time.

During this initial learning phase:

• cost per result (CPA) is extremely volatile and, on average, much higher than usual
• display frequency swings unpredictably
• the algorithm needs a minimum number of conversions to come out of the learning phase and settle down

According to Meta's official guidelines on delivery optimisation, an ad set typically needs around 50 optimisation events (for example, 50 purchases or 50 add-to-carts) within a week to exit the learning phase and stabilise message delivery — a threshold well worth double-checking on the Meta Business Help Centre when you actually publish, since the platforms update it fairly often. For Google Ads and Performance Max campaigns, the machine-learning systems typically need one to two weeks to process conversion signals and start optimising bids efficiently.

Now picture the scene.

It's Black Friday week. Between 23 and 27 November, every brand on the planet — from multi-million-euro multinationals right down to the clothes shop round your corner — is pouring huge sums into Meta and Google's ad auctions. Competition for people's attention rockets. As a result, CPMs (cost per thousand impressions) climb sharply compared to a normal Tuesday in October. If you decide to publish and launch your ads on 20 or 22 November, here's the perfect disaster that unfolds:

1. Your campaign enters the learning phase exactly when auctions are at their most expensive all year
2. The algorithm will burn through your budget at a terrifying rate trying to work out who to show the ads to, paying top whack for every single impression
3. Because you're not hitting the weekly conversions needed, thanks to the high costs, your campaign gets stuck in "limited learning" status

Result: you'll have spent a big chunk of your budget for a fraction of the results you could've got with a much smaller budget, spread out over the right timeframe.

Those who know how the system works do the exact opposite. They warm the engine up a month early. Launching promotional campaigns or ad sets at the end of October or the first days of November, on a reduced budget, lets the algorithm finish the learning phase while the ad auctions are still relatively quiet and CPMs are low. By the time 27 November comes round, the machine's already calibrated: the algorithm already knows who your ideal buyer is, the pixel has already tracked purchase intent, and the only thing left to do is bump up the daily budget on campaigns that are already running efficiently.

You're not improvising: you're putting your foot down on a machine that's already got a warm engine.

If you want to get into the operational detail of setting these campaigns up platform by platform, I covered that specifically in Meta Ads 2026: the practical guide for SMEs who don't want to burn their budget and in Google Ads for SMEs: how not to burn your budget in 30 days.

A true story, with a real mistake

A note for the reader: what follows is a composite case, built by combining real, recurring dynamics I've seen across several clients in the Campanian furniture and design sector — not a single identifiable client. The technical mechanics (shared hosting, no stress test, database timeouts) are exactly what I see repeated every single year, in November, with different business owners.

November 2024. A Campanian company producing furniture and interior design pieces — I'll call it Aura Design — decides to have a proper go at its first "big" Black Friday e-commerce push. They'd produced a special line of metal-and-wood mirrors and coffee tables.

The owner, blinded by enthusiasm and convinced that digital was just a matter of pressing a button, refused to start preparing in September. "We've got the products in stock," he'd say, "it'll take us ten minutes to knock 30% off on the site and fire off a few thousand euros of Instagram ads."

We get to Thursday 21 November 2024. Less than twenty-four hours before the Black Friday weekend, the makeshift agency they'd hired uploads the ad campaigns and activates the discount codes on the WooCommerce e-commerce site they'd had built on the cheap a few years earlier. The site sat on an £18-a-month shared hosting plan, never updated or tested in three years, with no caching system and no CDN serving the images. Nobody had checked the MySQL database could take the load, nobody had optimised product image sizes — they'd been uploaded in high resolution straight off the camera, several megabytes each.

Friday 22 November, 9.15pm. The ad campaigns start pushing traffic. Hundreds of users click the ads at once and land on the product page for the discounted coffee table.

The inevitable happens. The server can't handle the concurrent load. The database hits a connection timeout. Users' screens stop loading images and show a chilling white error message: 504 Gateway Timeout.

Over the following four hours — right in the middle of Friday-night impulse-buy peak — the site remains only intermittently reachable, with checkout pages taking over twenty seconds to load. The basket freezes up while processing card payments.

I can still remember the owner's phone call at 10.40pm that night. Gone was the usual confidence from the pitch meetings. There was just that embarrassed, desperate silence of someone watching a whole quarter's work go up in smoke. I could hear his fingers tapping nervously on the keyboard in the background as he tried, uselessly, to refresh his own site.

The final numbers from that disaster were a proper bloodbath: thousands of euros of ad budget burnt in four days driving people to a broken site, hundreds of baskets abandoned because of system errors, dozens of emails and DMs from customers convinced they'd fallen for a scam or a "fake discount" set up purely to collect clicks without delivering anything.

Aura Design's problem wasn't the product. The products were solid, the price was competitive, the photos were polished. The failure came down purely to a lack of properly tested infrastructure upstream, and the assumption that technology just magically works without ever being put under pressure at the right time. A simple server stress test run at the start of October, plus a migration plan onto proper infrastructure, would've made the difference between a record profit and a heavy loss once you factor in direct costs and lost revenue.

I've tackled the theme of infrastructure buckling exactly when you need it most from another angle too, in A backup isn't a precaution — it's your site's fourth wall — worth reading alongside this one.

Real cases from the big brands (and what they teach those without those budgets)

If you think forward planning is just something we consultants bang on about to make clients' lives harder, take a look at how the biggest commercial organisations on the planet operate — and how, sometimes, they fail.

The good — John Lewis (UK).

The British retail giant has run its Christmas advert every single year since 2007, without a break. In recent years, it's landed reliably between the start and middle of November — 8 November 2018, 4 November 2021, 9 November 2023, 14 November 2024, 4 November 2025 — with an almost ceremonial regularity that's become part of the British collective ritual of "Christmas starting" in itself. It's no accident that this punctuality has become as anticipated a media event as the advert itself: it means that, every single year for eighteen years running, production, media planning and budget have been locked in well ahead of launch, not decided on in autumn. It's not the size of the budget — which for a campaign of this scale has historically run into millions of pounds between production and media planning — that makes the difference: it's the culture of preparation time. The same principle, scaled proportionally, applies to an SME in Caserta, London or Manchester alike.

The bad — Target, United States, Black Friday 2013.

This is a textbook case of the principle that "security doesn't get tested in November." Between 27 November and 15 December 2013, right in the thick of the Christmas shopping season and starting from Black Friday, malware installed on Target's US point-of-sale systems let attackers lift roughly 40 million credit and debit card accounts' data. Target confirmed the breach publicly on 19 December 2013, after the news had already broken the day before and after the company had been notified by the authorities on 12 December. In January 2014, it also emerged that personal data belonging to up to 70 million customers had been stolen. The breach led to direct costs, litigation, investigations, falling sales and serious damage to brand trust, becoming one of the most-cited cases of what happens when systems aren't properly prepared for the year's peak commercial exposure.

The problem of discounts that aren't really discounts.

Every year, the British organisation Which? looks into how genuine Black Friday deals actually are. In its 2022 investigation, covering 214 offers from seven major retailers — including Amazon, Argos, Currys and John Lewis — Which? found that 183 products, 86% of the total, had been available at the same price or cheaper at least once in the six months before Black Friday. In other words, only around one deal in seven appeared to represent a genuine saving against recent price history, though this figure isn't in itself a legal ruling on unlawfulness.

Which? has published follow-up analyses since: in its investigation into Black Friday 2023, released in 2024, 92% of the 227 offers examined turned out to have been available at the same price or cheaper at other points in the year, outside the Black Friday fortnight. This isn't just a reputational or commercial issue: within the EU, Directive 2019/2161 — known as the Omnibus Directive — requires that any announced price reduction reference the lowest price the seller charged in the previous 30 days. In Italy, this was transposed into law via Legislative Decree 26/2023, applicable to price-reduction announcements from 1 July 2023. So you can't just freely invent an inflated reference price to make a discount look bigger than it is.

This regulatory detail is a nasty trap for unprepared SMEs. If you decide in mid-November to discount a product for Black Friday, but you'd already changed its price or run other last-minute promotions in the preceding 30 days, you could find yourself legally unable to advertise the discount percentage you wanted, risking a complaint from the Italian Competition and Market Authority (AGCM). Even from a legal standpoint, time is on your side: only by locking down your catalogue and price lists in September can you guarantee full compliance and maximum persuasive punch for your November discounts.

Berto vs Bobby: two speeds, same calendar

If there's one thing my double professional life between Italy and the UK has taught me to see with absolute clarity, it's the deep cultural difference in how the Italian and British markets approach the Q4 commercial calendar.

In the UK, Black Friday is a ritual perfectly transplanted from American culture and by now thoroughly woven into the British social and economic fabric. Bobby, in Manchester or London, operates within a system where the British consumer plans their Christmas shopping to line up with Black Friday. The UK market also has Boxing Day (26 December) on top, kicking off a second wave of sales before the year's out. For Bobby, the October-to-December quarter isn't a string of separate events — it's one single promotional block with two peaks: late November (Black Friday and Cyber Monday) and late December (Boxing Day and winter sales).

In Italy, the situation is radically different, and looking at it through American-marketing-tinted glasses is the biggest mistake any consultant can make.

Berto isn't lazy. Berto is viscerally, and justifiably, wary. His wariness comes from real experience: back in 2012 he was stitched up by a so-called web agency that promised him the earth with an e-commerce site that never actually worked, and ever since he's eyed anything smacking of forced importation with healthy Campanian scepticism. To Berto, Black Friday often feels like a passing fad that risks devaluing his products and undermining genuine Italian commercial tradition: the January winter sales.

In the Italian context, the January sales still carry huge cultural, social and economic weight. Italian consumers know the "real" sales land in shops in January, and they tend to split their spending: some in November for gifts or online shopping, a good chunk kept back for early January purchases at local shops or for clothing.

Telling a small business owner in Caserta or Portici to "just do what they do in London" is a shallow bit of advice. Italy's calendar shouldn't be a photocopy of the English one: it needs to manage two distinct seasons without letting them short-circuit each other.

• In November (Black Friday): focus on frequency and exclusivity. You don't need to flog the whole catalogue — fewer options, pushed harder, sell more, as I've already written about the marketing of subtraction. What you need is a surgical selection of loss-leader products, stock to clear, or purpose-built bundles.
• In December (Christmas): the lever isn't price any more, it's delivery guarantee. In Italy, people buying in December will happily pay full price as long as they're absolutely certain the parcel lands under the tree before Christmas Eve evening.
• In January (sales): focus on restocking and clearing collections, tapping into the traditional Italian spending habit in shops and online.

The job of communications isn't to turn Berto into a London trader — it's to help him understand that, to dominate his own local market, he needs to move his pieces with the same lead time as Bobby. The tactics change, but the physics of time stays exactly the same: slow planning destroys margin, whatever language you explain it in.

If you want the full-year view, not just Q4, I covered that in The calendar Berto's never made (and Bobby does by instinct): a guide to seasonal marketing for SMEs — this article is the deep dive into the trickiest quarter of the lot.

Operational checklist: what to do, week by week

You don't lose Black Friday on 27 November

There's a mistaken idea doing the rounds in office corridors and in the chats of plenty of amateur marketing consultants: the idea that Black Friday is a lottery, a lucky day where all you need to do is flick the advertising switch, make a bit of noise on social media, and wait for the payment notifications to rain down on your phone like it's New Year's Eve.

It doesn't work like that. It never has.

Marketing isn't a magic trick: it's an engineering process — logistical, psychological and financial — applied to time.

You don't win Black Friday on 27 November. You win it in mid-September, on that exact day you decide whether to leave your commercial calendar rolled up with its red elastic band stuck in an office drawer, or whether to have the nerve to unroll it on the table, look the deadlines square in the face, and start working backwards.

If you wait until November to decide what to discount, you'll end up chasing the market. You'll have rushed photos, copy written at two in the morning, a server that buckles at the first spike in visits, an advertising algorithm draining your margins in an endless learning phase, and a list of unhappy customers to explain to why their parcel didn't arrive in time for the holidays.

But if you decide to take the advantage of time now — while everyone else is still slowly sleeping off the summer holidays — the exact opposite happens. Come November, you won't need to run. You won't need to shout to be heard over the noise. Your machine will already be warmed up, tested, clean, and sat on the starting line. All you'll need to do is pull the lever and let your numbers run.

Berto looked again at the calendar unrolled on his table in Santa Maria Capua Vetere. Then he pulled a thick black marker pen out of the pot, found the box for 15 September, and drew a dark, heavy circle around it.

"Umbè," he said, before finishing his now-cold coffee, "ring the office. From today, Campania's calendar is open too."

The real question isn't what you'll do on 27 November. There's only one real question: is your 2026 calendar still stuck under an elastic band in a drawer, or have you already unrolled it on the table?