The clearest warning sign is a proposal that arrives without anyone asking about your
conversion rate, your margin, or your inventory position. DSP amplifies whatever your product
page already does, so a proposal built without those three numbers is selling reach rather
than proposing an outcome.
The short version
- No prerequisite questions means no diagnosis. The biggest warning sign.
- A large non-negotiable minimum is a business model, not a media plan.
- Unexplained view-through windows manufacture the result.
- Creative treated as an afterthought. DSP lives on creative.
- No stop conditions. Softer measurement means campaigns run on narrative.
The seven warning signs
I run Flapen with 50 operators managing about 70 brands. DSP is in scope for every
client, which is exactly why the proposals I see from elsewhere are worth describing.
| # | Warning sign | What it usually means |
|---|---|---|
| 1 | No questions about conversion, margin, inventory | Reach is being sold, not an outcome |
| 2 | Opens with awareness audiences | Skipping the cheap demand entirely |
| 3 | Large fixed minimum | Their economics, not your test |
| 4 | View-through window unstated | The number can be shaped later |
| 5 | Reuses existing product photography | Creative not resourced |
| 6 | Reports only in-platform metrics | Nothing outside DSP has to move |
| 7 | No stop conditions | Runs until you cancel it |
1. No prerequisite questions
DSP is a traffic instrument, and traffic instruments amplify what the detail page already
does. If conversion is weak, DSP buys more visitors for a page that does not convert. If
margin is thin, it accelerates losses. If inventory cannot absorb the volume, it buys a
stockout.
A serious partner asks for conversion rate, landed cost, and inventory cover before proposing
anything. A proposal built without them is a media plan, not a recommendation.
2. Awareness first
The cheapest customer is one already searching for your category, and the second cheapest is
one who viewed your detail page and left. Awareness audiences are the most expensive per
outcome, which is why they belong last.
A proposal opening with in-market and lifestyle audiences has skipped two cheaper rungs.
3. The fixed minimum
Minimums exist and are sometimes legitimate. What matters is whether the partner can design a
proportionate first test regardless.
Retargeting your own detail page visitors is a small, warm, cheap starting campaign. If a
proposal cannot accommodate that because of a monthly floor, the floor is describing their
business rather than your product.
4 and 6. Measurement that cannot lose
Two related patterns. An unstated attribution window means the result can be presented
generously later. In-platform-only reporting means nothing outside the DSP dashboard has to
improve for the campaign to be declared successful.
Ask which number outside DSP should move. Blended cost of customer acquisition, organic share
of revenue, or new-to-brand purchases are all acceptable answers. Silence is not.
5. Creative as an afterthought
DSP is a creative format. Targeting decides who sees it and creative decides whether anything
happens, and repurposed product photography is a materially weaker version of the format.
Ask who produces display and video assets, whether it is in-house, and how quickly a second
variant can be tested. Subcontracted creative on a format that depends on iteration is a real
constraint.
7. No stop conditions
DSP underperformance is unusually easy to explain away, because measurement is softer and the
judgment window is longer. That combination lets a campaign run for two quarters on
narrative.
Agree in advance what would trigger stopping, in the same way you would for a product:
defined metrics, a defined window, and a review date.
What a good proposal contains
A small first test, usually retargeting, at a budget proportionate to your account. A stated
attribution window with reasoning. Named creative deliverables and who produces them. One
number outside DSP that should move. A review date and stop conditions. And an honest note on
what DSP will not fix.
What most agencies will not tell you
DSP is a comfortable thing to sell. Higher budgets, softer attribution, longer judgment
windows, and a narrative about brand building that is true and conveniently
unfalsifiable in the short term.
None of that makes DSP a bad format. It makes it the format where the gap between a capable
partner and a reseller is widest, and where the proposal itself is the best available
evidence of which one you are talking to.
The other thing: if your sponsored search campaigns still have unharvested intent at your
current spend, that demand is cheaper than anything DSP will find. An agency proposing DSP
while your Sponsored Products campaigns are budget-capped has not looked at your account.
Related answers
- How to evaluate Amazon DSP capabilities
- Amazon DSP vs PPC which to use
- What to use for Amazon DSP if current partner lags
- Agencies that specialize in Amazon DSP creative
- Hiring an Amazon agency: the complete guide
We audit conversion and inventory before proposing any traffic spend. Flapen.
