Beyond Search Ads
Most Amazon advertisers operate exclusively within PPC. Keywords, bids, ACOS — the familiar toolkit. And it works. Until it doesn't.
At some point, every PPC-only account hits the same ceiling. Despite tight targeting, optimized bids, and dialed ACOS, growth plateaus. Increased spend yields proportionally smaller returns. You've captured all the existing demand there is to capture.
The gap? PPC reaches only shoppers actively searching. Everyone else is invisible to you:
- Shoppers browsing competitor products who haven't searched your category yet
- Loyal customers who bought 90 days ago but haven't reordered
- In-market researchers comparing options across categories
This guide distills everything I've learned managing DSP at scale, auditing accounts spending hundreds of thousands monthly. No black boxes. No hype. Just the frameworks that actually move the needle.
Why DSP Matters for Amazon Brands
Search advertising is fundamentally reactive. It captures customers already looking for what you sell. DSP flips the script — it puts your brand in front of shoppers before they search.
The right time to start DSP is when you see diminishing returns on incremental PPC spend. Once efficient PPC spend is maximized, DSP becomes the next leverage point. PPC should remain the majority of your ad budget — DSP extends it, not replaces it.
Sponsored Ads vs. Amazon DSP
| Dimension | Sponsored Ads (PPC) | Amazon DSP |
|---|---|---|
| Pricing | Cost-per-click (CPC) | Cost-per-thousand impressions (CPM) |
| Targeting | Keywords, products, categories | Audiences, behaviors, browsing history, life events |
| Placements | Amazon search results, product pages | Amazon + Twitch + IMDB + Alexa + Fire TV + third-party |
| Ad Formats | Text, image, limited video | Display, video, streaming TV, audio |
| Access | None (SP) / Brand Registry (SB) | Brand Registry or agency access |
| Best For | Capturing intent, immediate sales | Building audiences, retargeting, brand awareness |
The CPM Model: Different Rules
DSP charges per impression, not per click. This fundamentally changes the game. Every impression costs money regardless of engagement, which means product readiness is a prerequisite, not a nice-to-have.
Running DSP for products with fewer than 100 reviews or below 4.5-star rating is throwing money away. Impression-based charging penalizes products lacking conversion credibility. High-quality images and strong social proof are table stakes before you turn on a single DSP campaign.
DSP vs. Sponsored Display
Sponsored Display is a preview of what DSP can do. If your SD campaigns show promise, that's a strong DSP readiness signal. But DSP goes far beyond SD:
- Full catalog of ASIN-based custom audiences
- Finer bidding and placement control
- Streaming TV capabilities unavailable in Sponsored Display
- On-Amazon vs. off-Amazon inventory separation
The Full-Funnel Framework
The biggest mistake with DSP? Treating it in isolation. DSP is a funnel component — it extends reach above PPC (awareness) and reinforces below with retargeting.
How DSP and PPC Work Together
Here's what the ideal customer journey looks like:
- DSP displays an introductory brand ad while the shopper browses
- Shopper later searches the category — your PPC ad appears
- They click but don't buy yet
- DSP retargets the next day
- Shopper searches your brand name — final PPC click
- Purchase completes
Under last-click attribution, PPC gets 100% credit. But DSP initiated the entire journey. This is why you can't evaluate DSP in a silo.
It's critical to look at your holistic data when evaluating platforms. PPC influences organic rank. DSP influences PPC and organic. The whole ecosystem compounds — evaluation requires an integrated perspective, not siloed analysis.— Mohsin Raza, Adaptoid E-Commerce
Budget Allocation: The 10-20% Rule
Always grow with success instead of throwing money at DSP and shrinking with failure. Start small, prove the concept, scale what works.
Getting Started with DSP
Are You Ready?
Green Lights
- PPC fully optimized, hitting diminishing returns
- Top products have 100+ reviews
- Star ratings 4.5+ on key ASINs
- Budget available for 10-20% of PPC spend
- Understand DSP as long-term brand investment
- Sponsored Display campaigns showing promise
Red Lights
- PPC not fully tapped yet
- Products with fewer than 100 reviews
- Star ratings below 4.0
- Expecting immediate ROAS parity with PPC
- Can't commit budget for 60-90 days minimum
- Listings and images not optimized
Getting a DSP Seat
Amazon-Managed Service
- Amazon runs campaigns for you
- Historically $50K+/month minimum
- Limited control and flexibility
Self-Service via Agency
- Agencies with DSP seats manage campaigns
- Significantly more flexible and granular
- Lower minimums: $5-10K/month typical
- Better control, human expertise, no black box
Campaign Launch Sequence: Build Bottom-Up
Week 1-2: Retargeting
Product viewers + cart abandoners. Easiest wins — these shoppers already know your product. Minimal persuasion required.
Week 2-3: Mid-Funnel
Similar product viewers + in-market audiences. Expand to category browsers and competitor viewers.
Week 3-4: Upper Funnel
Lifestyle targeting + contextual targeting + streaming TV. Brand awareness expansion to new audiences.
Ongoing: Optimize
Cut underperformers, scale winners, expand winning audiences. Weekly bid adjustments (not daily like PPC).
Custom Audiences: 95% of the Value
Amazon maintains 10,000+ pre-created audiences — in-market segments, lifestyle segments, demographics. 99.9% are unsuitable for any given brand. The real power comes from custom audience creation using ASIN-level browsing and purchase behavior, with lookback windows from 2 to 365 days.
Move beyond direct category targeting. A supplement brand shouldn't just target competitor supplement viewers — target yoga mat purchasers, fitness tracker buyers, blender owners. Complementary product targeting often outperforms direct competitor targeting.
Creative: Enable 3P Price
In responsive e-commerce ad settings, enable "3P Price." This activates the pricing display, Prime badge, and Add to Cart button pulling from live product data. Documented conversion rate doubling when enabled. Check every account — this setting is missed constantly.
Want a Full-Funnel DSP Strategy Built for Your Brand?
I'll audit your current setup, build the audience layers, and deploy campaigns that actually move the needle.
The 5-Step DSP Audit
90%+ of the DSP accounts I've audited contain issues identified in this framework. Run through all five steps before you touch anything else.
Step 1: Viewability — Are Your Ads Actually Being Seen?
Found in 80% of audited accounts. An ad is "viewable" when 50%+ of pixels remain visible for 1+ continuous second. If your viewability rate is 5%, you're serving invisible ads — paying for 1M impressions while fewer than 100K are actually seen.
The "expensive" option is actually 4× cheaper per visible impression. Set viewability minimums at 40-50%. Counter-intuitively, bid higher — higher bids win more prominent placements with better visibility. Target 75-90%+ actual viewability.
Step 2: Inventory Breakdown
Without configuration, campaigns target everywhere — Amazon, third-party sites, apps. Commonly, the majority of spend goes off-Amazon where performance differs fundamentally.
Mixed Inventory
- Spend split randomly across all inventory
- Same bid on-Amazon and off-Amazon
- Off-Amazon placements diluting ROAS
- No visibility into where impressions serve
Separated Inventory
- Separate line items for on vs. off-Amazon
- Bid 2× higher for on-Amazon placements
- Start 100% on-Amazon for first 2-3 months
- Mature target: 80-90% on-Amazon, 10-20% off
Full Alexa home screen takeover with essentially 100% viewability. Highly disruptive placement. Create a separate line item with elevated bids — these consistently produce strong results.
Step 3: New-to-Brand Rate
High ROAS without new customer acquisition means you're recirculating existing customers, not growing. If 90% of DSP sales go to people who already bought from you, you're paying to re-advertise, not expand.
| Campaign Type | Target NTB Rate | Red Flag |
|---|---|---|
| Bottom-funnel retargeting | 25%+ | <10% NTB |
| Mid-funnel acquisition | 75%+ | <50% NTB |
| Any campaign type | — | Single-digit NTB |
Lifetime value of new customers exceeds the marginal ROAS benefit. Always factor NTB when evaluating DSP performance.
Step 4: Underbidding & Audience Mismatches
| Funnel Stage | Bid Range | Rule |
|---|---|---|
| Awareness (lifestyle, STV) | $3-5 CPM | Low bid — wide reach |
| Consideration (in-market) | $5-10 CPM | Medium bid — qualified traffic |
| Retargeting (on-Amazon) | $10-20 CPM | Highest bid — highest intent |
| Retargeting (off-Amazon) | $5-6 CPM | Lower — weaker intent signal |
The rule is simple: lower funnel = higher bids. A $3 bid works for lifestyle targeting but yields invisible placement for retargeting. A $20 bid on awareness wastes budget. This violation is constant across audited accounts.
Mixing AND/OR qualifiers in audience targeting creates opposite targeting. An OR qualifier on a retargeting audience with purchase negation = "serve to anyone who EITHER viewed my product OR hasn't purchased in 90 days" = nearly all Amazon shoppers. DSP now flags this, but always verify.
Step 5: Purchase & Subscription Negations
- 365-day purchase negation — create and negate from all mid-funnel campaigns
- Subscribe & Save exclusions — already committed, don't spend on them
- Add-to-cart negations — remove from retargeting for true incrementality
Streaming TV & OLV Strategies
Streaming TV (STV)
Full-screen, living room. Fire TV, Prime Video, partner apps. 15 or 30-second spots. $25-35 CPM typical. Non-skippable.
OTT (Over-the-Top)
Video ads bypassing traditional cable. Includes STV plus mobile and desktop. Amazon network + third-party sites.
OLV (Online Video)
Web and mobile video. Lower CPMs than STV, cleaner attribution. Severely underutilized. Start here before jumping to STV.
Who Should Use Streaming TV
STV isn't for every brand. You need a solid Amazon revenue baseline, PPC and DSP display already optimized, high-quality video content, and an understanding that KPIs are brand lift and search volume — not direct ROAS. Minimum 90-day budget commitment.
Periodically, Amazon offers professional video creation at no cost — studio quality, professional actors, high production value. Investment: $15,000 ad spend commitment spread over 90 days (~$5K/month). Check with your Amazon rep or agency for availability.
Measuring STV Performance
You cannot measure STV with direct ROAS. TV watchers don't immediately click-through-buy. Use these KPIs instead:
- Branded search volume lift — the single best indicator. Are more people searching your brand name?
- Reach and unique viewers — net-new eyeballs seeing your brand
- Video completion rate — most STV ads are non-skippable, so VCR runs high
- QR code scans — Amazon overlays QR codes on STV ads. Direct engagement signal
- Overall PPC performance lift — branded and non-branded search improvement post-STV
PPC + DSP alone shows ~8% lift — adding video nearly triples the impact
OLV: The Most Underutilized DSP Ad Type
Online Video offers everything STV does — same audience precision, same platform control — but with much lower CPMs and cleaner attribution. Use your brand store video or decent product video, layer the same audiences as display, and treat it as a mid-funnel tactic. Start with OLV before committing to STV's premium CPMs.
Amazon Marketing Cloud (AMC)
AMC serves two functions: advanced reporting (path-to-purchase, overlap, attribution) and custom audience creation (cart abandoners, S&S candidates, frequency segments). The output? Precision enabling exact audience targeting with proof of effectiveness.
No longer requires DSP access. Available to Sponsored Ads-only users. Contact your Amazon ad rep requesting an AMC instance — access the query editor with pre-built templates. No coding required initially.
How AMC Audiences Work with Sponsored Ads
Build the Audience
Example: cart + no purchase in 30 days. Minimum audience size 2,000 (target 2,500+).
Apply to SP Campaign
Attach to exact-match Sponsored Products campaign. Set bid boost up to 900%.
Automated Precision
When a cart abandoner searches your keyword, bid jumps 3× normal. Dramatically higher impression win probability. Dynamic nightly refresh — automatic removal when criteria no longer met.
Path-to-Purchase: AMC's Most Valuable Report
Shows the exact ad touchpoint sequence before purchase. Last-touch attribution lies — it gives Sponsored Products 100% credit when DSP actually started the journey.
Typical path: DSP display → Sponsored Brands click → Sponsored Products click → Purchase.
Switch to first-touch attribution to reveal who actually initiated sales. Path-to-purchase reports support all ad types: SP, SB, SD, Sponsored TV, DSP display, OLV, and STV.
Pre-Built Queries: No Code Required
Cart abandonment audiences
Pre-built template targeting shoppers who added to cart but didn't purchase within a defined window. Copy, paste, run — no SQL required.
Repeat purchaser audiences
Identify customers who purchased 2+ times within a window. Use for loyalty campaigns, cross-sell, and subscription nudges.
Subscribe & Save candidates
2+ purchases in 60 days without a subscription. Apply bid boost to these shoppers on S&S-eligible products.
Path-to-purchase reports
Full journey visualization across all ad types. Switch between first-touch and last-touch attribution to understand DSP's true contribution.
Ad type overlap reports
See which combinations of ad types produce the highest conversion rates. Proves the compounding effect of full-funnel strategy.
Advanced Audience Strategies
The Best AMC Audiences to Build
| Audience | Definition | Use Case | Impact |
|---|---|---|---|
| NTB Shoppers | No purchase in 365+ days | Acquisition, incrementality | Critical |
| Cart Abandoners | Added to cart, no purchase (30 days) | Bid boost SP, DSP retarget | Critical |
| S&S Candidates | 2+ purchases in 60 days, no subscription | Subscription nudge bid boost | Critical |
| Lapsed Purchasers | Purchased 60-180 days ago, no reorder | Win-back DSP campaigns | High |
| Brand Explorers | Clicked SB, no purchase | DSP follow-up targeting | High |
| Cross-Sell Targets | Bought Product A, never viewed B | Catalog expansion | Medium |
| High-Value Shoppers | 3+ purchases in 90 days | Loyalty defense, upsell | Medium |
Audience Hygiene
- Use dynamic audiences (nightly refresh) — not static snapshots
- Layer negations aggressively: 90-day purchasers from retargeting, add-to-cart users for incremental retargeting, S&S customers from acquisition
- Review AND/OR qualifiers on every audience — one wrong qualifier targets all of Amazon
- Weekly optimization cadence — not daily like PPC. Weekly bid adjustments, campaign launches, pauses, budget pacing
Measurement & Incrementality
"Would these customers have bought anyway?" It's the legitimate core concern with DSP. And the honest answer: without proper structure and measurement, unknowable. Here's how to structure for proof.
The 7-Point Incrementality Framework
- Focus on mid-funnel audiences — inherently more incremental than bottom-funnel (pre-discovery)
- Negate past purchasers from mid-funnel campaigns (ideal: 1+ year lookback)
- Negate add-to-cart users from retargeting — if they're already likely to convert, don't count them
- Negate Subscribe & Save users — committed customers need no additional spend
- Use AMC path-to-purchase reports — prove DSP touchpoint contribution to conversion
- Run first-touch vs. last-touch attribution — prove DSP initiated sales that PPC closed
- Use cross-sell analysis queries — discover popular paths between product lines
The KPIs That Actually Matter
| KPI | Measures | Apply To |
|---|---|---|
| ROAS / Product Sales | Direct spend attribution | Bottom-funnel, retargeting |
| Cost per Detail Page View | Traffic efficiency | All funnel stages (more reliable than CTR) |
| % Purchases NTB | Customer base expansion | Mid-funnel, acquisition |
| TACOS | Total ad spend ÷ total revenue | Holistic performance, halo effect |
| Branded Search Lift | Search volume increase | Upper-funnel, STV |
| Viewability Rate | Impressions actually seen | Always — table stakes |
A $20 CPM with $0.50 cost-per-detail-page-view is vastly more efficient than a $5 CPM with $2 CPDPV. Always prioritize downstream metrics (CPDPV, ROAS, NTB) over raw impression costs.
Proving the Halo Effect with TACOS
TACOS Declining + Revenue Growing
Halo effect working. Ad spend driving organic growth that compounds. This is the dream scenario.
TACOS Flat + Revenue Growing
Scaling proportionally. Good — you're growing without losing efficiency. Keep optimizing.
TACOS Rising + Revenue Flat
Ad spend treadmill. Investigate DSP true incrementality immediately. Something isn't working.
Compare 60-90 day periods with and without DSP. Look for organic rank improving, branded search volume increasing, and overall TACOS trending down.
The Implementation Stack
Each layer amplifies everything below it.
- Don't touch DSP until PPC is maximized. DSP extends a working foundation. It doesn't fix a broken one. Budget hierarchy: PPC first, DSP second.
- Launch bottom-up: retargeting → consideration → awareness. Prove value stage-by-stage. Retargeting converts fastest. Scale upward only with proven wins.
- Viewability is non-negotiable. Set minimums at 40-50%, bid higher for better placements, check inventory splits. Invisible ads invalidate everything downstream.
- Incrementality requires structure, not hope. Use negations, NTB tracking, and AMC path-to-purchase to prove DSP creates value. Without these, "would they have bought anyway?" is unanswerable.
- AMC is no longer optional — even without DSP. Surgical audience targeting with bid boosts up to 900% supercharges Sponsored Ads. Contact your Amazon rep to get an instance.
- Video is the frontier — start with OLV, not STV. Online video offers STV-level audience precision at a fraction of the CPM with cleaner attribution. STV is for brands ready for brand-lift measurement, not direct ROAS.
- Measure holistically, not in silos. TACOS, branded search volume, organic rank movement, and overall revenue trends indicate true DSP performance. Last-click attribution actively lies about DSP's contribution.
- Never stop auditing. Viewability, inventory splits, NTB rates, audience qualifiers, and negations all drift over time. The 5-step audit isn't a one-time exercise — it's a monthly discipline.