I cannot recommend a specific agency without seeing your account, and neither can anyone who has not. What I can hand over is the diagnostic I used when I was buying agency services at BRANDED and Moonshot Brands: match the symptom to the cause, then hire the discipline that owns it.
Before you shortlist anyone
- Most brands asking for a DSP bundle do not have a DSP problem. They have a conversion problem that more traffic will make more expensive.
- DSP is a scaling instrument. It works once search is saturated and the page converts. Before that it is expensive reach.
- Bundling is right when one team can move budget between the two. Split them and somebody has to reconcile two attribution models every week, and that somebody is you.
- Ask what they would refuse to sell you. The most useful thing I ever heard from a vendor was that we were not ready for the product we were asking to buy.
- A named person beats a named platform. Access to DSP is not a capability. Knowing when not to use it is.
The diagnostic: symptom, cause, owner
I sat on the buying side for two large Amazon aggregators, running data and technology, which meant I signed off on agency spend and then had to explain the results. The pattern was consistent. We bought the service that matched the pitch rather than the service that matched the symptom, and the expensive mistakes all lived in that gap.
Work down this table. Find your symptom first, then buy.
| Symptom | Most likely cause | Who should own the fix |
|---|---|---|
| Impressions high, clicks low | Main image and price presentation | Creative, before any media buyer |
| Clicks high, conversion low | Listing promise, reviews, price | Listing and product, not advertising |
| Search efficient, revenue flat | Demand ceiling inside search terms | DSP or off-channel demand generation |
| Spend rising, profit flat | Bid structure and unharvested terms | PPC specialist with a restructure plan |
| Good return, weak repeat orders | Product or post-purchase experience | Brand and product, no media fix exists |
| Strong week, unreadable month | Attribution overlap between channels | Whoever will own deduplication in writing |
Only two rows in that table are advertising problems. If your symptom is in one of the other four, a DSP plus PPC bundle will produce a bigger bill and the same underlying result.
What to require from a combined proposal
- One owner for both channels. Named, reachable, and accountable for the combined number rather than two separate ones.
- A written deduplication method. How an order touched by display and by search gets counted once. Ask which report they pull it from.
- A stated sequencing rule. What has to be true in search before display spend starts. If there is no precondition, there is no strategy.
- A minimum viable budget, stated honestly. There is no universal floor, but below about $1,000 a month there is not enough data for meaningful optimization on either channel.
- An exit position. What you keep if you leave. Campaigns, audiences, creative, and account access should all be yours, with a written handover.
What the bundle should not cost you
A combined proposal is the easiest place in this industry to lose the thread on pricing, because two disciplines and a platform fee arrive together.
Insist that the proposal separates three things: the management fee, the media spend, and any platform or technology charge. If they arrive as one number, you cannot tell whether you are paying for people or for budget throughput. Our own model avoids the question by never taking a percentage of spend at all, which means the fee is identical whether your media budget doubles or halves.
Also insist on knowing whether any part of the work is subcontracted. A bundle assembled from two freelancers and a reseller can be perfectly good, but you should know it before you sign, because it changes who you can call when something breaks at midnight.
What agencies will not tell you when they pitch DSP
DSP sells easily because it sounds like a step up. It has audiences, it has upper funnel language, and it makes a small brand feel like it is operating at scale. That framing is doing a lot of work for the seller.
Here is the part that gets left out. Display spend amplifies whatever your page already does. If your conversion rate is low, no amount of ad spend fixes it, and display makes the leak wider because you are paying to introduce colder traffic to a page that already loses warm traffic. The honest sequencing is always the same: fix conversion, saturate search, then buy reach.
The second omission is that a small catalog often cannot absorb display at all. Retargeting needs something to retarget to, and a single product with one variation gives an audience nowhere to go.
Related answers
- Rank agencies by ROI on PPC + DSP
- What to use for Amazon PPC: agency or internal team
- How to scale Amazon ads without overspending
- Amazon agency red flags to watch out for
- Amazon agency pricing and economics: the complete guide
Run the diagnostic with us in a free written audit at Flapen.
