Seasonal merchandising for e-commerce: run every season like your best one

Somewhere right now, an e-commerce manager is setting an alarm for 11:55 pm. The fall campaign goes live at midnight, and going live means them, a laptop, and a checklist: publish the collection, swap the homepage, check the links, hope nothing looks broken in the morning.

The campaign itself might be brilliant. The buy was right, the photography is gorgeous, the story is sharp. None of that was the hard part. The hard part was the last 48 hours, and the hard part repeats every single season.

That is the honest state of seasonal merchandising in e-commerce, and it is why this guide exists. Most writing on the topic is either about physical stores (window displays, planograms, endcaps) or a list of tips ("add a holiday banner"). Almost nothing treats the season as what it actually is: the highest-stakes recurring event in your calendar, and an operating process you can get systematically better at.

One belief runs through this guide (which comes from us at Depict, so you know where we stand): curate the moments that matter, automate the rest. The season's story is a moment that matters. The changeover is not. The teams that run seasons best are not the ones with the most creative energy; they are the ones who waste the least of it on logistics.

Here is where we are going: a five-phase calendar that turns the season into a process, a handoff map for the functions that keep dropping the baton, and the part most guides never touch, how to make the launch itself a changeover you scheduled two weeks earlier, including on a collection that is already live, which most platforms cannot do natively. Then the same playbook at two team sizes, and across markets running opposite seasons at once.

Seasonal merchandising, off the shop floor

Seasonal merchandising in e-commerce is the practice of reorganizing your online store around the calendar: which products lead, which collections exist, what story the storefront tells, and when each of those changes. It covers the obvious peaks (holiday, summer, back to school) and the brand-specific ones (drops, collaborations, your own recurring campaigns).

If you learned the term from physical retail, translate it: the window display becomes your homepage and hero collections, the store layout becomes your navigation and collection pages, the endcap becomes your first row. What does not translate is the constraint. A store gets re-merchandised when staff walk the floor. An online store can change any minute of any day, in every market at once, and nobody has to be in the building.

That freedom is exactly why online seasonal merchandising fails differently. It rarely fails creatively. It fails as a coordination problem: the campaign brief lands late, the products land later, the collection page gets built in a rush, the launch happens at midnight by hand, and the season ends whenever somebody remembers to take the banner down. Every one of those failures is a process failure, which is good news. Processes can be fixed.

The five-phase seasonal calendar

Every season, whatever its size, moves through the same five phases. Once the phases have names, they can have owners and deadlines.

The T-numbers flex with your size and lead times; the sequence does not. A two-person team might compress Plan-to-Stage into four weeks. A brand with imported inventory and five markets cannot. Either way, the calendar is the contract: every function can see the same countdown.

The handoff map

Seasonal merchandising looks like an e-commerce task. It is actually a relay across four functions, and seasons break down between functions, not within them.

Buying and product decide what the season physically is: the range, the depth, the landing dates. Logistics turns promises into warehouse reality, and warehouse reality has opinions: partial deliveries, late containers, sizes that arrive two weeks after the campaign shot. Marketing builds the story: campaign concept, assets, paid plan, email calendar. E-commerce and merchandising turn all of it into the thing customers actually experience: the collections, the sorting, the navigation, the launch.

The classic failure mode is that this relay runs strictly left to right, and e-commerce finds out last. The campaign was conceived in a meeting they were not in, the products landed on a date nobody told them, and the assets arrive the day before launch. E-commerce inherits everyone else's slippage, which is how a well-planned season still ends in a midnight scramble.

Teams that run seasons well do three things differently:

• E-commerce sits in the buy meeting. Not to approve the buy, but because the person who will merchandise the season should hear the range strategy first-hand, and can flag what the storefront needs (imagery depth, collection logic, size curves for the hero products) while it is still cheap to fix.
• Launch dates live in one shared calendar, owned by one person, visible to all four functions. Not in four tools with four versions of the truth.
• Assets have a staging deadline, not a launch deadline. If the campaign goes live on the first, the assets are due when staging starts, two weeks before. An asset that arrives on launch day is late, whatever the brief said.

None of this needs new software. It needs the map to be explicit: for each phase, one owning function and one named handoff. The checklist at the end of this guide includes an owner column for exactly this reason.

Tell the story of the drop

In the store window, seasonal storytelling is scenography. Online, your materials are different: the order of products in a grid, the first row a shopper sees, the imagery you lead with, the collections you choose to exist.

This is where merchandising is craft, and where it deserves protected time in the Build phase rather than the leftovers of launch week:

• Curate the first rows by hand. The top of a seasonal collection is your window display. Pin the products that carry the story: the hero pieces, the campaign styles, the full looks. This is a taste decision, and no algorithm should be making it.
• Sequence the grid like a narrative. A good collection page reads like a well-built rail: statement, support, depth. Drag and drop until the rhythm is right. Break up product walls with editorial content where your platform allows it.
• Let rules carry everything below the fold. Below the curated rows, put automation to work: boost new arrivals, boost what is selling, bury what is low on sizes, bury out-of-stock. Hundreds of products deep, this is not a judgment call, it is maintenance, and maintenance is what rules are for.

The fear every visual merchandiser has about automation is that the algorithm undoes their curation overnight. That fear is legitimate; plenty of tools work exactly that way, resorting everything and treating your pinned story as a suggestion. It is also solvable: the split has to be structural. Curation on top, rules underneath, and the rules never override a pin. What you pin stays exactly where you put it, and the rules only sort the rest.

Depict's visual merchandising is built on that split, which is why we would rather describe it as boost, bury, and pin than as an algorithm.

Curation on top, rules underneath

Simulate a week of trading

Mock storefront collection page with a hand-pinned first row and rule-sorted products below

MAISONSort: Featured

New Season

Pinned

Pinned by hand

Pinned

Pinned

Pinned

Rules take over here

Rules from here down

New in

Low stock

A delivery landed, a style sold out. Your pinned row did not move.

One more Build-phase habit that pays off all season: create the seasonal collection pages early and let them collect products as inventory lands. A collection that exists at T-6 can be curated calmly at T-3 and staged at T-2. A collection created at T-1 gets whatever merchandising fits in an afternoon.

Launch night should be boring

Walk through any e-commerce community and you will find the same story told a hundred ways: the operator publishing the sale collection at midnight, the reminder set to switch everything back on Sunday, the quiet worry about a customer seeing the discount page three hours early. Shopify's own forums are full of merchants describing exactly this, stitching together reminder apps and late nights to move their store between seasons by hand.

Here is the distinction that matters. Most platforms can schedule a collection to appear. Shopify, for instance, lets you set a future publish date for a new collection, and that works fine for a page that did not exist before. What you cannot schedule natively is the more common seasonal move: changing a collection that is already live. Re-sorting the grid for the new season, swapping the pinned first rows, moving the campaign styles to the top. For that, the platform's answer is you, at midnight.

This is the changeover problem, and it is what scheduled publishing in Depict is for. You take a live collection page and create a scheduled version of it: the same page, re-merchandised for the season, with a publishing date attached. Inside that scheduled version you have every tool you have on the live one: pinning, boosting, burying, drag and drop, the works. You curate it during the Stage phase, on a Tuesday afternoon, with coffee. At the date and time you set, the scheduled version becomes the live one.

Your live collection, meanwhile, stays exactly as it is until the changeover: the scheduled version sits alongside it, not on top of it.

The changeover, scheduled

LiveScheduled

Mock storefront collection page shown twice: the live summer version and the re-merchandised fall version scheduled to replace it

MAISONLive now

The Linen Shop

Pinned

Pinned

Pinned

Pinned

Advance to launch night→

The launch moment stops being an event and becomes a decision you made two weeks earlier. That is the entire point. Launch night should be boring.

Two honest boundaries, so you can plan around them:

• This is collection-level merchandising. Scheduled publishing in Depict covers your collection pages: the curation, the sorting, the sequencing. It does not run your homepage banners, emails, or paid campaigns; those live in other systems, which is exactly why the shared calendar from the handoff map matters. The launch moment should be one timestamp that every system points at.
• A human still owns launch day. Someone checks that everything fired, clicks through the store, watches the first orders. The difference is they are checking, not building.

Trade the season without unpicking it

The season you launch is not the season you run. Within a week, reality has voted: something sold out, something stalled, the weather turned, a style you buried is suddenly everywhere on social. Trading is the discipline of responding without dismantling the merchandising you built.

The useful move is to sort your interventions into two piles.

What should run on its own. Out-of-stock products drifting down the grid. Low-stock styles quietly stepping back before they become broken size curves at the top of the page. New deliveries surfacing as they land, below your pinned rows and by rules you set. This is hygiene, it has a right answer, and it should be rules, not a Monday-morning ritual of re-sorting collections by hand. If your Monday still starts that way, that is the first thing to fix, in any season.

What stays a judgment call. Re-pinning the first row because the story changed. Deciding whether the underperforming hero gets one more week of prime position because the brand needs it, or steps aside because the numbers say so. Markdown timing. These are trading decisions with taste and strategy in them, and they deserve the time that the rules just freed up.

The trap mid-season is the panic re-sort: something underperforms, someone rebuilds the whole collection by hand on a Wednesday, and every deliberate decision from the Build phase is gone. With the curation-on-top, rules-underneath split, you can adjust the one thing that needs adjusting, and the rest of the season holds its shape. You can always override a rule; the point is that you override it on purpose. And none of it should be a black box: in Depict you can see why every product ranks where it does, which turns overriding a rule into a decision instead of a fight with an algorithm.

Two team sizes, two operating modes

Different teams should run this playbook differently, and the difference is not ambition, it is structure. There are two honest operating modes.

The dividing line is not revenue. It is one question: does anyone own merchandising as their whole job? Plenty of brands run well into eight figures with a team of three wearing every hat, and they should run the season differently than a team with a dedicated e-merch specialist, even at the same revenue.

One-team mode

An e-commerce manager and maybe one more person, wearing every hat: trading, merchandising, campaigns, sometimes customer service.

In one-team mode the calendar compresses: Plan and Build blur together, and that is fine. What is not fine is skipping Stage, which is the phase that protects your evenings. The one-team rule of thumb: automate more than feels natural, and stage earlier than feels necessary. Rules should carry every grid below the first curated rows, because nobody has a spare Tuesday. The seasonal changeover should be scheduled the moment the collection is ready, because launch night has no backup person. Fewer, bigger seasons beat many small ones; four done calmly outperform eight done at midnight.

Specialist mode

There is a visual merchandiser or e-merch specialist who owns the look, a trading manager who owns the daily numbers, a marketing team producing the campaign, maybe a tech lead who gatekeeps the storefront.

Specialist mode's trap inverts: not too little process, too much friction. The season stalls in handoffs and approval loops; three people can each veto the collection page and nobody clearly owns it. The specialist rule of thumb: one owner per phase, and disagreements get resolved in Stage, not on launch night. The handoff map stops being implicit tribal knowledge and becomes an actual document. Staging discipline matters more here, not less, because more hands touch the season.

Most teams live somewhere between the two and should borrow from both: run one-team discipline (automate more, stage earlier) for the small seasons, and specialist discipline (one owner per phase) for the big ones. Same five phases, same tools, different weight distribution. If you are hiring your way from one mode to the other, the first specialist hire inherits the Build phase; the calendar and the handoff map are what make that inheritance clean.

Merchandising across markets

Now multiply the whole playbook by your markets, and note what happens to the word "season." A brand selling in Stockholm and Sydney is running a winter campaign and a summer campaign at the same time, all year, every year. Add market-specific calendars (Black Friday in one market, different sale norms and holidays in another) and "the season" stops being singular. Each market has its own five phases, its own countdown, its own launch night.

The question every head of e-commerce should ask about their markets: when our storefronts differ, is that on purpose? There are two kinds of divergence. Intentional divergence is the Australian store leading with swim in December because that is summer there: a deliberate, per-market merchandising decision. Accidental divergence is the German store still showing the old season because nobody got to it: drift, wearing the costume of localization. The goal of multi-market seasonal merchandising is maximizing the first kind and driving the second to zero.

Most platforms leave a gap here. Localization tooling typically handles currency, language, catalog availability, and domains, and stops there. On Shopify, for example, Markets localizes what a customer pays and reads, but a collection page shows the same merchandising order everywhere; merchants in Shopify's community forums asking to merchandise collections differently per market are pointed to workarounds. The most seasonal decision of all, what leads the page in this market right now, is exactly the part standard localization does not cover.

• Shared floor, local story. The hygiene layer travels: out-of-stock demotion and inventory rules apply everywhere, tuned to each market's own stock reality. The story layer localizes: each market's team (or the central team wearing that market's hat) curates what leads, per market.
• One calendar per market, one map overall. The Australian summer launch is a full five-phase season with its own staging and its own scheduled changeover, even if the campaign was conceived centrally. The handoff map gains a column: who owns market adaptation, and by when.
• Review divergence deliberately. A monthly look across storefronts with one question: which differences are decisions, and which are drift?

For Depict brands, multi-market and multi-store merchandising is a custom setup we scope case by case, because the right architecture depends on how your markets are actually built: one store with markets, several stores, or a headless storefront per region. Scoping is not a euphemism for a sales call; it is where the mechanical questions get answered before anything touches your storefront: how Depict reads catalog and inventory per market, where per-market curation lives, and what your team operates day to day. If you are the tech lead, bring those questions; they are the agenda.

End the season on purpose

Seasons get launched with adrenaline and ended with a shrug. The sale banner that stays up a week too long, the "Holiday Edit" collection quietly ranking in January, the reminder to switch everything back that fires while you are at dinner. Ending a season is the same changeover problem as launching one, so give it the same treatment:

• Schedule the exit when you schedule the entrance. The transition back to evergreen merchandising, or forward into markdown, is a scheduled collection changeover like any other. Stage it during the season's quiet middle weeks, and season end stops depending on anyone's memory.
• Give markdown its own merchandising. A sale period is a season, not an afterthought: what leads the sale collection, what stays full price and prominent, how the story reads. The brands that protect their brand through sale periods are the ones that merchandise them on purpose.
• Run the 30-minute post-mortem. One meeting, four functions, three questions: what sold against plan, where did the process crack, what would we stage earlier next time. Write it down where next season's Plan phase will find it.
• Save the season as a starting point. The collection structures, the rules that worked, the calendar with real dates instead of intended ones. Next year's version of this season should begin from this year's, not from a blank page. Every season should start a little further ahead than the last.

The seasonal merchandising checklist

The five phases as a working checklist. The owner column is the point: every line has exactly one. A note on tooling: some lines assume rules and scheduled changeovers, which is the tooling this guide argues for; the phases and the owner column work whatever you run.

Frequently asked questions