Look for in-house production and a stated iteration cadence, not a portfolio. DSP creative
lives or dies on how fast a second version can be tested, and subcontracted production makes
that slow and expensive enough that the first version becomes permanent.
The short version
- In-house production beats portfolio quality. Iteration speed is the variable.
- Ask how fast a second variant can be tested. Days, not a new quote.
- Cold audiences need different creative from detail page imagery.
- Ask what they test: hook, format, message, or all three.
- Creative decides DSP outcomes more than targeting does.
What DSP creative actually has to do
I run Flapen with 50 operators managing about 70 brands, with product photography,
videography, 3D rendering, and infographic design produced in-house through a Dubai studio.
DSP audiences are colder than search audiences. Someone who typed your category into Amazon
has declared intent. Someone seeing a display ad has not, which changes what the creative has
to accomplish.
| Surface | Audience state | Creative job |
|---|---|---|
| Sponsored Products | Actively searching | Show the product clearly |
| DSP retargeting | Viewed and left | Remove the specific hesitation |
| DSP in-market | Category interest | Establish relevance fast |
| DSP awareness | No context | Earn attention before anything else |
Reusing detail page photography across all four is the most common failure. A white-background
product shot works in search results, where the buyer is already looking for the product. It
does almost nothing to a cold audience scrolling past.
The questions that identify real creative capability
- Who produces the assets? In-house or subcontracted, and where.
- How fast can a second variant be tested? Days is capable. Weeks is a quote cycle.
- What formats do you produce? Static, animated, and video have different economics.
- What do you test first? Hook, format, or message.
- Show me two variants of the same concept and explain why the winner won.
Question two is the whole thing. DSP performance comes from iterating toward what works, and
an agency that has to raise a new brief and a new quote for each variant will test once and
call it done.
Question five is the best test of judgment. Anyone can produce assets. Explaining why one
outperformed another means they have been reading results rather than delivering files.
Why iteration matters more than initial quality
A good first creative and a fast iteration loop beats a brilliant first creative and no loop,
almost every time.
Nobody predicts which hook lands with a cold audience for a specific product in a specific
category. You find it by testing. Three mediocre variants tested in a month teach you more
than one excellent variant shipped and left running for a quarter.
This is also why in-house production matters as a constraint rather than as a credential.
It is not that in-house creative is inherently better. It is that when a test suggests
changing the first three seconds of a video, an in-house team does it this week.
What to look for beyond assets
Ask how creative connects to the rest of the account.
The strongest DSP creative usually comes from the same research that drives the listing:
competitor negative reviews, the rating gap, and the specific objection that stops purchases.
An agency producing DSP creative without reference to why people currently do not buy is
producing decoration.
Also ask whether the same team produces your listing images. Consistency between what someone
sees in an ad and what they find on your detail page reduces the drop-off between the two, and
splitting the work across providers reliably breaks it.
What most agencies will not tell you
Creative is frequently the least resourced part of a DSP engagement and the one that decides
most of the outcome. Targeting is easier to talk about, easier to report on, and easier to
adjust, so it dominates the conversation.
Ask what proportion of the engagement is creative production versus campaign management. If
creative is an afterthought in the proposal, it will be an afterthought in the work.
The second thing: portfolio quality is a weak signal here. Portfolios show the best work an
agency has produced, often for large brands with large budgets. What predicts your outcome is
how quickly they will produce the fourth variant of something that is not working, which no
portfolio can show you.
Related answers
- How to evaluate Amazon DSP capabilities
- Warning signs in Amazon DSP proposals
- Amazon DSP vs PPC which to use
- Recommended agency for Amazon storefront design
- Hiring an Amazon agency: the complete guide
Ask how fast we can test a second variant. Our studio is in-house at Flapen.
