DSP stands for demand side platform, and Amazon operates one for display advertising. It buys placements against audiences rather than against the keywords a shopper types. The budget question is economic, so decide it on your own session volume, margin, and retargetable audience, never on a competitor's review count.
The short version
- DSP stands for demand side platform. Amazon runs its own, and it buys audiences where sponsored ads buy search terms.
- The decision is arithmetic before it is technology. Work out your cost per incremental order before anyone opens a proposal.
- 90+ data points, not a review count. That is the standard we hold a market to, and a display budget answers to it too.
- Sponsored search holds the first claim on the money. Display makes sense once search is efficient at your mature target.
- Ask for the current terms in writing. Amazon sets access and minimums, those terms change, and no proposal replaces the written number.
What DSP stands for, and what the money buys
Sponsored ads answer a search term, and display answers a person. That single difference is the mechanism, and every economic consequence below follows from it.
A shopper on a results page has already decided to buy in your category. A shopper reached by display has decided nothing, so the persuasion is yours to fund.
At Flapen, 50 operators run about 70 brands by hand, and our own engineers build the advertising tools they work in. So I treat display as a capital allocation question, since the hard part is proving the orders would not have arrived without it.
Display repairs nothing that is already broken on your detail page. It multiplies whatever your listing does today, in both directions, and it cannot create margin the unit economics never had.
The arithmetic that decides a display budget
Run this on paper before you accept any proposal. Every line uses numbers you already own.
- Write down contribution margin per unit. Price minus landed cost, Amazon fees, and the cost of returns, which needs to sit under 8% for the margin to survive.
- Turn that margin into a percentage of price. Contribution margin before ad spend, divided by price, is your break-even ACoS.
- Count the audience you can address. Take detail page sessions over the last 30 days, minus the shoppers who already bought.
- Set the baseline before anything runs. Record total orders, branded search volume, and organic rank for the four weeks preceding the flight.
- Price the incremental order, not the attributed one. Divide total flight cost by the orders above that baseline, then hold it against break-even.
- Read the outcome on a fixed window. Our kill criteria run 60 to 90 days, and a display test earns the same treatment.
Step five is where most proposals die, because the attributed number flatters the flight while the incremental number is what your bank balance reads.
| Cost line | The number to write down | Who owns it |
|---|---|---|
| Media | Total committed across the flight, not a monthly average | You |
| Management | A flat monthly figure, or the percentage and what it becomes at triple the budget | Your provider |
| Creative | Cost per refresh, and who keeps the files at the end | Your designer |
| Inventory | Cash tied up in the units the flight is supposed to sell | You |
| Measurement | The sales you forgo during a holdout, priced honestly | Both of you |
Five lines sit in that table, and a proposal quoting only the first one has quoted a fifth of the cost.
The data that decides it, and the data that does not
We read 90+ data points before committing to a market, and review count is among the weakest of them. Reviews describe last year, and a display budget is a bet on next quarter.
The ones that answer this question are category growth trajectory, return rate, conversion rate per ASIN, the trajectory of your acquisition cost, and how fast detail page sessions refresh. Inventory cover belongs there too, because a stockout mid-flight burns the media and the rank it bought.
Once a product is live, four signals decide whether it scales, gets fixed, or gets killed: rating trend, return rate, conversion rate, and cost of customer acquisition trajectory. Read all four before you buy reach. Display can move volume, and it cannot move a rating trend or a return rate.
Paid is one of five traffic channels, beside organic, promotions, influencer and creator programs, and off-channel. Most sellers run two of the five, so the money a display flight consumes often buys more units in a channel nobody has switched on.
So ask any candidate what they analyze besides reviews and sales volume. A team that names return rate, session refresh, and margin per ASIN is doing research. A team that names your competitors' review counts is doing a pitch.
What a display proposal will not tell you
Four things get left out of this conversation, and on a bad week that includes ours.
Start with the fee arithmetic. A fee set as a percentage of media pays its owner more every time the budget rises, while your contribution margin per unit does not move at all.
Triple the budget and the fee triples. Your margin per unit is identical, so the increase comes out of your profit unless incremental orders climb faster than the spend.
| What triples with the budget | The provider paid on spend | You |
|---|---|---|
| Media | Fee triples | Cash out triples |
| Attributed sales | Reported as the result | Partly orders you already had |
| Incremental orders | Rarely quoted at all | The only line that pays you back |
The second omission is view-through credit. A report can count the order of a shopper who ignored the ad, then searched your brand name a week later. Without a baseline or a holdout, you pay for that order twice.
The third is that the honest answer for many brands is no display this quarter, and that answer earns nobody a commission.
The fourth omission is ours to own. We charge a flat monthly fee with every service included and no commission on your spend, so telling you to skip display costs us nothing. Hold that against us, and if our answer on your account is a display plan with no incremental measurement in it, do not hire us.
Related answers
- ACoS meaning
- PPC on Amazon
- Amazon marketplace expansion from UAE to USA
- B2B Amazon marketplace strategy in Germany
- Amazon account measurement and audits: the complete guide
This week, at no cost, run the first three steps above on your best selling ASIN: contribution margin per unit, break-even ACoS, and 30 days of detail page sessions minus buyers. If that last number is small, the display question is answered for this quarter.
Send those three numbers over and get a written audit back inside 48 hours at no charge from Flapen.
