eCommerce PPC management services for budgets split across platforms that all claim the same sale
Run ads in four places and you will receive four reports, each confidently crediting itself with revenue the others also counted. Added together they routinely exceed everything you actually sold. Nobody is lying. Each platform can only see the part of the journey it touched, and none of them is incentivised to mention the rest.
We manage the whole spend against one number that cannot be double counted: what the business actually banked, and what it cost to get it.
One budget across every platform, blended numbers that cannot be double counted, and decisions made on profit after fees rather than reported ROAS.
What are eCommerce PPC management services?
eCommerce PPC management services run paid advertising across every place a product business can buy traffic: marketplace ads inside Amazon, eBay or Etsy, Google Search and Shopping, and paid social. The distinguishing work is not managing any one of them, it is deciding how a single budget should be divided between them when each reports its own success in a currency the others do not recognise.
Where you have one dominant channel, go to the specialist page instead. Amazon PPC management goes deeper into Amazon's auction than this page will, and Google Shopping deeper into feeds and bidding. This page is for sellers spending meaningfully in three or more places at once, where the allocation question has become bigger than any individual account.
It also differs from general PPC management in what it optimises towards. Lead generation optimises for cost per lead. Product businesses have to optimise for profit per order after platform fees, fulfilment and returns, which are large enough to turn a good ROAS into a loss.
Every platform reports a version of events where it was responsible
A customer sees your product in a paid social post, searches your brand on Google a day later, clicks a Shopping ad, and buys. Meta records a conversion because it showed the ad that started it. Google records a conversion because it delivered the click that ended it. Both are accurate descriptions of what that platform observed, and both describe the same single order.
On marketplaces the same effect appears in a different costume. Attribution windows credit ads for purchases made days later by shoppers who may well have bought anyway, and the platform that sets those windows is the platform being paid. There is no independent audit available to you, which is not a scandal so much as a structural fact that has to be planned around.
The practical consequence is that reallocating budget on the strength of any single dashboard is reallocating on an overlapping number. Sellers pour money into whichever platform reports best, which is usually the one with the most generous attribution rather than the one producing the most incremental revenue. The two are unrelated.
Attribution answers a different question from the one you are asking. It tells you which platform touched a sale, not whether that sale would have happened anyway. Those are only the same number when advertising reaches people who would otherwise never have found you, and that is precisely the situation no dashboard is able to identify for you.
Claims sum past 100%
When the reported revenue across platforms exceeds what you actually sold, the excess is the overlap. It is a useful diagnostic and almost nobody checks it.
Branded search flatters everybody
People who already decided to buy still have to arrive somehow. Ads that intercept them look excellent and often add nothing that would not have happened anyway.
ROAS ignores the fee stack
A four to one return can still lose money once referral fees, fulfilment, returns and the cost of goods are subtracted. On marketplaces the fees are large enough to invert the answer.
Marketplaces mark their own homework
The company setting the attribution window is the company selling the advertising. Treat their reporting as directional, and judge outcomes on total profit.
Blended numbers, because they cannot be double counted
If per-platform reporting overlaps, the honest alternative is to stop asking each platform how it did and start asking the business. Total advertising spend against total revenue is a crude measure, and its crudeness is exactly why it is trustworthy: there is no attribution model inside it to be generous.
Blended cost of sale
Every dollar of advertising against every dollar of revenue. One number, no overlap, and the only one that moves when the business genuinely improves.
New customer cost
Advertising divided by first-time buyers. Harsher and more useful, since existing customers were not won by the ad that intercepted them.
Deliberate holdouts
Turn a channel off for a defined period and watch total revenue rather than that channel's dashboard. Uncomfortable, occasionally revelatory, and the closest thing to proof available.
Contribution per order
Revenue minus goods, platform fees, fulfilment, returns and advertising. The number that decides whether more volume helps or hurts.
Branded and unbranded split
Separated in every report, because mixing them lets brand demand you already earned disguise the performance of everything else.
Judged over months
Product businesses have seasons and repeat purchase cycles. A fortnight of data on a seasonal catalogue is noise presented as insight.
Platform dashboards still get used, for what they are genuinely good at: telling you which campaign, keyword or product is doing better than its neighbour inside the same account. That is a comparison within one attribution model, which is fair. Comparing across models is not.
How the budget should actually be divided
The instinct is to split the budget by which platform reports the best return. A better approach is to divide it by the job each platform does, because they are not substitutes for one another and treating them as interchangeable produces a portfolio that only harvests demand it did not create.
Marketplace advertising captures people already shopping with a wallet open. It converts well and it is almost entirely harvest: those buyers were going to buy something in that category today. Search captures existing demand for what you sell, with branded search harvesting demand you already earned and unbranded search reaching people who do not yet know you. Paid social creates demand among people who were not looking, which converts worse on first contact and is the only line item that grows the pool everything else is harvesting from.
A portfolio that is entirely harvest looks superb in reporting and stops growing, because nothing is refilling the top. A portfolio that is entirely demand creation looks poor in reporting and burns cash. The split between them is a strategic decision about growth rate and risk appetite, and it should be made deliberately with you rather than emerging from whichever dashboard was most flattering last month.
How our eCommerce PPC engagements run
Establish real economics
Contribution per order after goods, platform fees, fulfilment and returns. Until this exists, no target return on ad spend can be set honestly.
Measure the overlap
Add up what every platform claims and compare it to actual revenue. The gap tells you how much of your reporting is describing the same orders twice.
Allocate by job, not by dashboard
Harvest and demand creation budgeted separately, with branded search sized deliberately rather than allowed to absorb whatever it wants.
Run, hold out, adjust
Day to day management inside each account, with periodic holdout tests on the channels whose contribution is genuinely uncertain.
The first month usually produces one uncomfortable finding. Most often it is that branded search is being credited with revenue that would have arrived regardless, and that the budget behind it is better spent elsewhere.
What our eCommerce PPC management services include
Unit economics model
Contribution per order per product after every cost, which sets the target return each channel has to clear rather than an industry rule of thumb.
Blended reporting
One report covering total spend, total revenue, blended cost of sale and new customer cost, with each platform's own claim shown beside it for context.
Marketplace advertising
Campaign structure, targeting and bids inside Amazon, eBay, Etsy or Walmart, managed against margin rather than against the platform's preferred metric.
Search and Shopping
Google Search and Shopping campaigns fed by a clean product feed, with branded and unbranded separated so their performance never hides inside one number.
Paid social
The demand creation part of the portfolio, budgeted as an investment in future harvest rather than judged on the same day-one return as marketplace ads.
Holdout testing
Structured pauses on channels whose incremental contribution is unclear, measured on total revenue, agreed with you before anything is switched off.
Product level decisions
Advertising concentrated on items that can carry it after fees, and withdrawn from those that cannot, which is often the fastest improvement available.
Seasonal budgeting
Spend raised into the periods that carry the year and pulled out of the flat ones, planned in advance rather than reacted to.
An agreed stop rule
The point at which we say a channel is not worth funding for your catalogue, set in advance so the recommendation is credible when it arrives.
When more ad spend is the wrong move
The listings do not convert
Paid traffic to a page that fails to sell buys the same failure faster. Fix conversion first, on the marketplace or on your own store.
Margin cannot carry a click
On low-value items the fee stack plus advertising frequently exceeds contribution. Raise order value through bundles before raising budget.
You are only harvesting
If everything is branded search and marketplace ads, more budget buys diminishing returns on demand you already have. The constraint is awareness, not spend.
Stock will not support it
Advertising into a thin stock position produces cancellations, and on marketplaces cancellations cost visibility that outlasts the campaign.
Work out your real contribution per order first with our eCommerce profit margin calculator. That number, not an industry benchmark, is what any target return on ad spend should be derived from.
Paid media alongside the rest
Amazon PPC Management
Deeper on the auction, match types and placement inside the platform where most product ad money goes.
Fix my Amazon ads →Google Shopping Ads
Feed quality, bidding and structure on the channel that runs on your product data rather than your keywords.
Get my products in Shopping →eCommerce CRO
The cheapest way to lower cost per order is usually to convert more of the traffic you already pay for.
Raise my conversion rate →eCommerce SEO
The traffic you do not rent. Slower to build and it does not stop the day you pause the budget.
Grow my organic traffic →Marketplace Management
Stock, pricing and profit by channel, which is where the numbers behind any allocation decision come from.
Run my channels properly →eCommerce Management
The pillar service, for sellers who want the operation and the advertising looked at together.
See the whole picture →How much do eCommerce PPC management services cost?
A flat monthly fee, not a percentage of spend. Charging a share of your budget pays us more for spending more of your money, which is precisely the wrong incentive on a page arguing that most sellers should spend less in at least one place.
Audit and economics
$900 to $2,500 once. Contribution per order modelled properly, attribution overlap measured, and a recommended allocation with the reasoning shown.
Two to three channels
$1,200 to $2,800 a month. Day to day management, blended reporting and seasonal budgeting across your main platforms.
Full portfolio
$2,500 to $6,000 a month. Marketplace, search and social managed together, with holdout testing and product level allocation decisions.
Ad spend is paid to the platforms directly. Where the audit shows a channel cannot pay back on your catalogue, we will tell you to stop funding it, which is the one recommendation a percentage-of-spend agency has no reason to make.
Common questions about eCommerce PPC
Because each platform counts every sale it touched, and a customer often touches several before buying. Meta credits the ad that introduced you, Google credits the click that closed it, and the marketplace credits an ad seen days earlier. All three describe the same order. Adding the reports together always overstates, and the size of the overstatement is a useful measure of how much overlap sits in your reporting.
There is no universal figure, and any agency quoting one has not looked at your margins. The break-even return depends entirely on contribution per order after cost of goods, platform fees, fulfilment and returns. A business with wide margins can profit at two to one, while a marketplace seller with a heavy fee stack may need five to one on the same product. Calculate yours rather than adopting a benchmark from somebody else's catalogue.
Sometimes, and it is worth less than the reporting suggests. Branded search reliably shows an excellent return because it intercepts people who had already decided to buy. The genuine reasons to do it are defensive: stopping competitors appearing above you, or controlling the message when your organic result is not doing that job. Test it with a holdout and watch total revenue rather than the campaign's own numbers, which will always look wonderful.
Compare contribution per order rather than revenue, because marketplace fees change the answer substantially. Your own store usually keeps more per order but requires you to generate the demand yourself, whereas marketplaces bring demand and charge for it. Most sellers run both and use marketplaces for acquisition while working to make repeat purchases happen on their own store, where the economics and the customer relationship are better.
Long enough to cover a full purchase cycle, and longer for seasonal catalogues. Demand creation channels take longer to judge than harvest channels by their nature, because their effect appears in later searches and direct visits rather than in the click that led to the sale. Judging paid social on day-one attributed return will always conclude that it fails, which is a measurement artefact rather than a finding.
Not necessarily, and it is worth resisting the assumption that you do. Advertising is most valuable when it reaches people organic visibility cannot, or defends a position under attack. If you already rank well and sell well, additional spend often buys clicks you would have received free. That is a real risk on branded terms and on marketplaces where an ad appears directly above your own organic listing.
They work reasonably well and they optimise towards the platform's goal, which is revenue attributed to that platform rather than your profit after fees. Automation is genuinely good at finding demand and poor at knowing that one product loses money on every sale. The workable approach is using automation for discovery while controlling what it is allowed to advertise, which is a decision that has to come from your margin data.
Find out how much of your reporting is double counted
Give us access to your ad accounts and we will add up what every platform claims, compare it to what you actually banked, and show you the overlap. Then we will tell you which channel is genuinely adding revenue and which one is being paid for orders that were happening anyway.
Get my free ad spend review- Overlap measured
- Break-even return calculated
- Allocation you can defend