Ecommerce campaign

What a campaign is for

A campaign is a bounded push with one offer, one through-line, and one number attached to it. That last part is what separates a campaign from activity. If nobody can say what result would have made this worth running, the board will still fill up, the emails will still go out, and at the end there will be nothing to learn.

The one rule: decide the number before you decide the creative. The goal written into the brief is what makes every later argument settleable. Without it, "should we discount?" is a matter of taste. With it, it is arithmetic.

The order the decisions go in

Brief, then audience, then message. Reversing that order is the most common way a campaign quietly fails: someone has an idea for an email, the idea becomes the campaign, and the audience gets chosen afterwards to fit it.

• The offer comes before the copy. Copy can make a good offer land harder. It cannot rescue an offer nobody wants. If there is no real reason to buy now rather than in three weeks, name that as the problem instead of writing around it.
• A segment is only worth splitting out if it changes something. Two audience cards that would receive identical emails are one audience card. Split when the objection differs, the offer differs, or the channel differs.
• The through-line is one sentence, and every asset has to be recognizably the same campaign. Five emails making five different arguments are five campaigns sharing a send window.
• Pick the attribution window before the campaign runs. Seven-day click and one-day view is the common honest default on paid; last click for email. Choosing the window after the results are in is how a flat campaign becomes a success.

Failure modes worth naming out loud: discounting to hit a revenue number that a margin number would have vetoed; sending to the whole list because segmentation felt like extra work; shipping creative that tested well internally and has never been shown to a stranger; declaring victory on open rate.

Numbers to hold this against

Treat these as the shape of normal in ecommerce, not as targets. Your own last four campaigns are a better baseline than any published median.

• Open rate is directional now, not a measurement. Apple Mail Privacy Protection fires opens with no human involved, and depending on your list mix that can be a third to a half of what you see. Campaign opens commonly read between 25 and 40 percent with that inflation included. Trend it, never target it, and never call a test on it alone.
• Click rate on a broadcast campaign usually sits between 1 and 3 percent. Segmented sends run meaningfully above that; a whole-list blast usually below it.
• Revenue per recipient is the campaign metric worth arguing about, because it survives changes in list size. Typical ecommerce campaign sends land somewhere between 10 and 30 cents per recipient, and average order value plus send frequency move it more than copy does.
• A subject line test needs real volume to mean anything. Detecting a two point lift on a 30 percent open rate takes roughly 8,000 recipients per arm at conventional confidence. On a 20,000 person list, most subject line tests are coin flips being read as insight. Test offers and audiences instead: the effect sizes are larger and the reading is cheaper.
• Discount arithmetic, before the discount decision. At a 60 percent gross margin, a 20 percent discount needs about a 50 percent lift in units just to hold gross profit flat. At a 50 percent margin, the same discount needs the units to double. Run that sum on the brief's real margin before anyone types "20% off" into a subject line.
• Paid social prospecting for ecommerce commonly returns somewhere between 1.5 and 3 times ad spend, with retargeting higher on a much smaller pool. CPMs swing more by season than by creative, so never read a November CPM against a February one.

What has to be true before a card asks for approval

Email. One primary call to action, repeated, rather than five competing ones. A subject line a person would actually say, and a preheader that adds information instead of repeating the subject. The offer, its expiry, and its exclusions all visible without a second scroll. Alt text on every image, because a real share of recipients read with images blocked. Every link tagged, so the Measure phase has something to read. No image placeholder and no unresolved bracket surviving into approval.

Creative. The first frame carries the message with no sound and no caption. The claim in the ad matches the claim on the landing page, word for word where it can. The destination is the page for this offer, not the homepage.

Anything carrying a price or a claim. Price, dates, and terms checked against the brief rather than against memory.

A card that fails one of these is not ready for awaiting_approval, however good the writing is.

The seed cards are the first instance, not the ceiling

This board ships with five cards because five is the smallest arrangement that shows what each phase is for. A real campaign is far denser, and a thin board is usually a thin campaign.

Healthy counts for a two to four week push: the brief plus three to six research cards; two to four audience cards, each with its own objections; one plan and three to eight insight cards; in Create, one card for every asset that will actually ship, which is typically three to five emails and three to six creative variants rather than one of each; and in Measure, one result card per claim the brief made.

If a phase holds one card and the campaign is already underway, that phase is a heading rather than a phase.

The layout is one arrangement of many

The columns here run by production step: research, audience, strategy, create, launch, measure. That is the right default when a campaign has one offer and one window. It is not the only axis available. A multi-channel push often reads better arranged by channel, a seasonal one by wave, a multi-product one by product line. Rename the phases, add them, reorder them; the method survives the rearrangement. What does not survive is a campaign with no brief and no number.

When this is the wrong playbook

Reach for a specific playbook when one fits. A welcome flow, a winback, a launch, and a giveaway each have their own logic and their own failure modes, and this board will make you rebuild them from scratch. Skip it entirely when there is no offer, only a wish to be more visible: that is a positioning problem, and a campaign board will dress it up rather than solve it. And if the site converts its current traffic badly, more traffic is an expensive way to find that out.