Paid media · Ecommerce
Ecommerce ads judgedon contribution.
Brand search will make a blended ROAS look calm. The work is whether prospecting still clears margin after returns.
The short answer
What ecommerce paid media has to answer.
Ecommerce paid media is a contribution problem that happens to use ad accounts. Marketing efficiency ratio is total revenue over total marketing cost, once you have decided what counts as cost. On accounts where we cut prospecting that did not pay back inside 14 days, blended MER settled between 2.2x and 3.4x. That range is what those accounts did after the cut. It is not a target we write into a contract. Platform ROAS is a different picture. Meta prospecting has returned 1.6x while branded search has returned 6x. We do not average those into one number and call the account healthy. Brand is demand the store already earned. Prospecting has to stand on the margin of a new order, after product cost, shipping, returns, and fees.
Buying cycle
A first order is not a refill.
Someone buying a considered product will compare, leave, and come back on your name. Someone reordering a consumable already knows the brand. If both clicks share one ROAS cell, the refill and the brand query subsidize prospecting that is losing money on a new customer. We split first order from repeat before we talk about scaling.
The promo calendar sits inside that cycle. A sitewide discount week inflates platform ROAS and teaches the account to buy cheap revenue. Set the full-price new-customer target before the sale. When it ends, the budget comes back down. Returns arrive later, and a campaign that looked fine on day three can lose money once that rate is in.
Constraints
Margin, feed, and the checkout.
We do not run this without a margin. If finance cannot get close on product cost, shipping, fees, and returns, the first month is measurement, not a budget increase. A guess labeled as a guess is usable. A ROAS target copied from another brand is not. Creative that advertises a code the site will not honor does not ship.
The page can waste a clean account. On a separate store, checkout completion moved from 46% to 54% after the checkout itself changed. Bids did not do that. Shopping spend pointed at a checkout that drops orders rents the leak.
- New-customer orders stay apart from repeats, so email and brand search cannot hide inside prospecting.
- The product feed matches live price, availability, and variant. A disapproved SKU is merchandising, not a bid strategy.
- Promo codes in ads exist on the site the same day. Expired codes come out of the account.
- A category with a much higher return rate does not inherit the storewide cost per order.
Channels and KPIs
Google, Meta, and a number that can end a test.
Google carries brand search, non-brand search, and Shopping or Performance Max when the catalog and the margin can support the opacity. Meta carries prospecting and a smaller retargeting set with its own stop rule. Microsoft is added when those queries already convert and a second auction is worth the ops cost, not as a logo on the proposal.
We read MER for the business and platform ROAS by campaign type, side by side. Prospecting can be fine on a high-margin product and a loss on a thin one. Brand search does not bail it out. Those figures never collapse into one target.
- Cost per new order against contribution, not against last month's blended ROAS.
- Brand impression share, so a competitor on your name is visible.
- Prospecting payback inside the window you agreed, then a cut or a keep.
- Feed health: disapprovals, price mismatches, and products with spend and no margin.
Mistakes
One ROAS for three different jobs.
The usual failure is one Performance Max campaign, a branded search campaign nobody opens, and a Meta account optimized to all purchases, including view-through traffic from people who were going to buy anyway. The dashboard says the blend is fine. Contribution says the new customer was expensive. The second failure is scaling during the sale and leaving that budget in place when prices return to normal.
- No single ROAS target pasted across brand, Shopping, and prospecting.
- No prospecting budget increased because branded search returned 6x.
- No catalog ads for products that lose money after ads, shipping, and returns.
- No Performance Max left alone while the search terms and the product report go unread.
Example engagement
A specialty retailer with a known margin.
Specialty retailer, Shopify, contribution margin known after shipping and returns. The weekly report led with a blended ROAS. Branded search was most of the efficiency. Meta prospecting sat in the same paragraph, which made the blend look like a strategy.
We split brand, non-brand Shopping, and Meta prospecting so each had a stop rule. Prospecting that did not pay back inside 14 days came down. Across accounts where we made that kind of cut, blended MER settled between 2.2x and 3.4x. That range is not a promise for this store. On the split we refuse to merge, Meta prospecting returned 1.6x and branded search returned 6x. The second number does not repair the first.
2.2x to 3.4x
Blended MER across accounts after the cut, not this store's promise
1.6x vs 6x
Meta prospecting versus branded search, not averaged
The example uses anonymized results from a Quorum engagement. It is one account, not a benchmark you should budget against.
Questions
Ecommerce questions.
Should we pause brand search to make prospecting look better?
Usually no. Brand search is where people go after an ad, a referral, or a previous order. Pausing it so non-brand math looks cleaner is an accounting trick. We do cap what we are willing to pay for the brand term, and we watch competitors bidding on the name. The efficiency of brand is not a reason to hand prospecting the same target.
Do you require Performance Max?
No. It is useful when the feed is clean, the margin is known, and we can still see which products spend. If the campaign is a box that quietly buys the brand name, we stay on standard Shopping and search until the account can afford that opacity. A feed full of out-of-stock variants is not a reason to turn it on.
What do you need from finance before spend goes up?
A contribution view: product cost, typical shipping, payment fees, and a return rate that is at least directional by category. Perfect books are rare. We will not invent a margin to make a channel look viable, and we will not scale a category we cannot price.
How should email and SMS revenue show up in the ad report?
On their own line. A refill from a flow should not be poured back into the ad platform so prospecting can take credit. If you already run email, the paid review and the lifecycle review need the same definition of a new customer. Otherwise both meetings celebrate the same order.
Can you guarantee a ROAS?
No. We will agree on a contribution line and cut spend that misses it. A guaranteed ROAS is usually brand and retargeting in a costume. 6x on branded search and 1.6x on Meta prospecting stay in separate columns. We do not average them into a promise.
Does the platform of the store change the program?
The economics do not. The feed, the checkout, and the way an order is tagged do. Shopify is the common case. WooCommerce and BigCommerce are in scope when the catalog and the team can keep a product feed honest and a purchase event reliable.
Talk through ecommerce.
One conversation. A diagnostic. A plan you can kill if it is not the work.