Amazon PPC & Full-Funnel Ads

Part of the AMZ Operator Series (companion to IDS) | Built July 2026 | Distilling: Destaney Wishon (BTR Media) · Elizabeth Greene (Junglr) · Mina Elias (Trivium / PPC University) · Brent Zahradnik (AMZ Pathfinder) · Michael Facchin (Ad Badger / PPC Den) · Bradley Sutton (Helium 10) · Amazon Ads official docs · Pacvue & Ad Badger 2025–26 benchmark reports

AMZ 06 covered PPC as a rank weapon (CPR math, 1.5–2× launch bids, TACoS 30–50%). This module is everything after: ads as a permanent P&L line gliding to ≤20% TACoS by day 60–90 and falling.

---

#1. THE OPERATING FRAME — TACoS IS THE NORTH STAR

#1.1 The doctrine (Elias / Wishon framing)

ACoS = ad spend ÷ ad-attributed sales (campaign efficiency). TACoS = ad spend ÷ total sales (business health). Mina Elias's Trivium runs $700M+ of client revenue on contribution margin + TACoS reduction — not ROAS — because ACoS can be "optimized" into oblivion: cut every aggressive keyword, ACoS looks beautiful, organic rank decays, and six months later the account is smaller and more paid-dependent. Destaney Wishon's version: every ad dollar buys profit today or organic position tomorrow — TACoS is the only metric that sees both.

Rule: ACoS is a campaign dial. TACoS is the account verdict. Never sacrifice TACoS trend to make ACoS pretty.

#1.2 ACoS vs TACoS divergence — the interpretation table

Read the two lines together weekly:

#1.3 TACoS and organic:paid benchmarks by product age

Consistent with the AMZ 06 glide path (launch 30–50% → ≤20% by day 60–90):

A steadily falling TACoS is the single best evidence your ads are building an asset, not renting revenue.

---

#2. CAMPAIGN TAXONOMY — EVERY AD TYPE, WHEN IT EARNS BUDGET

Lifecycle allocation (share of total ad spend, directional):

Launch (0–2 mo):  SP 90–100%  ·  SB 0–10%   ·  SD 0–5%    ·  DSP/STV 0%
Growth (2–9 mo):  SP 70–85%   ·  SB 10–15%  ·  SD 5–10%   ·  DSP/STV 0–5%
Mature (9 mo+):   SP 55–75%   ·  SB 15–20%  ·  SD 10–15%  ·  DSP/STV 0–15%

SP is always the spine. Adding SB/SD before SP is efficient is decoration on a broken engine.

---

#3. ACCOUNT ARCHITECTURE — STRUCTURE IS STRATEGY

#3.1 Campaign structure doctrine (the 2026 consensus)

Elizabeth Greene's (Junglr) rule ended the SKAG-vs-lump war: structure by objective, not by match type or habit. A campaign trying to rank a keyword at 70% ACoS and defend a brand term at 8% ACoS "can't serve two masters" — the bid logic fights itself.

• Single-keyword campaigns (SKC): only for your 3–10 hero keywords — the terms that carry rank and revenue. Max control over budget, placements, and dayparting.
• Themed campaigns: everything else — one tight theme per campaign, one ad group, 5–15 closely related keywords sharing intent and economics. Pure account-wide SKAG is dead: thousands of campaigns thin the data and make optimization impossible.
• BTR Media's six objective buckets (Wishon): Awareness · Rank · Efficiency · Branded · Conquesting · Auto. Every campaign belongs to exactly one bucket, and each bucket has its own target ACoS. If you can't say which bucket a campaign is in, it shouldn't exist.

#3.2 Naming convention (non-negotiable — reports are useless without it)

[AdType] | [Objective] | [Match/Target] | [Product] | [Theme]
SP | RANK  | EXACT  | COLL-30  | collagen-powder-core
SP | AUTO  | CLOSE  | COLL-30  | discovery
SB | BRAND | VIDEO  | BRAND    | brand-defense
SD | EFF   | VIEWS30| COLL-30  | retarget-viewers
Filterable in any report by every token. Set once; never rename mid-flight.

#3.3 Portfolios and budget hierarchy

• Portfolio per product line (or per brand): portfolio-level budget caps are your guardrail against runaway spend; campaign budgets are sized so winners never cap before 8pm (AMZ 06 rule — a capped winner is rank you paid for and then refused).
• Budget priority order: Branded defense (small, fixed) → Rank (funded to plan) → Efficiency (scales with proof) → Conquest/Awareness (surplus only).

#3.4 The auto→manual harvesting loop (the engine of the whole system)

WEEKLY SEARCH TERM REPORT WORKFLOW
1. Pull search term report (or use tool), trailing 30 days
2. PROMOTE: search term with ≥2–3 orders AND ACoS ≤ target
   → add as EXACT in the matching manual campaign (set bid via §4)
   → add as NEGATIVE EXACT in the source (auto/broad/phrase) campaign
   Never promote without negating — otherwise you bid against yourself.
3. NEGATE: 0 orders AND clicks ≥ 2.5–3× expected clicks-per-order
   (expected = 1 ÷ CVR ≈ 10 clicks at 10% CVR → negate at ~25–30 clicks)
   Shortcut: spend ≥ 2× target cost-per-order with 0 orders → negative exact
4. Irrelevant terms (wrong product/use): negative PHRASE immediately, any spend
Cadence: weekly for accounts >$100/day spend; biweekly below.

#3.5 Placement multipliers and bidding modes

• Top-of-search converts 2–3× rest-of-search. Multipliers go up to +900%; the operator play is isolation: a dedicated ToS campaign for hero keywords (base bid low, ToS +50–150%) and a parallel campaign with ToS at 0% — clean data per placement. Check the placement report monthly; fund what wins.
• Dynamic bidding — down-only is the default, everywhere. You set the ceiling; Amazon can only cut. Up-and-down (Amazon may raise up to +100% at ToS) only on mature, consistently profitable campaigns with stable CVR — it hands Amazon your wallet on a placement where it grades its own homework. Fixed bids: rank-forcing tests and placement experiments only.

---

#4. BID MATH — THE FOUR FORMULAS THAT RUN THE ACCOUNT

#4.1 Breakeven ACoS = pre-ad contribution margin

Breakeven ACoS = (Price − COGS − referral − FBA fee − other variable) ÷ Price

Example ($29.99 product):
  29.99 − 6.20 (COGS) − 4.50 (referral 15%) − 6.10 (FBA) − 0.70 (other)
  = 12.49 contribution → Breakeven ACoS = 12.49 ÷ 29.99 = 41.6%
Below 41.6% ACoS an ad-attributed sale makes money; above, it loses.
Know this number per ASIN before touching a single bid.

#4.2 Target ACoS by lifecycle

Target ACoS = Breakeven ACoS × posture multiplier
  Launch/rank (0–2 mo):     1.0–1.5× breakeven  (≈ 40–65% — buying rank, per AMZ 06)
  Growth (2–9 mo):          0.6–0.8× breakeven  (≈ 25–35%)
  Profit/mature (9 mo+):    0.4–0.6× breakeven  (≈ 15–25%)
  Brand defense:            fixed low band       (≈ 5–15%)
Each BTR-style objective bucket gets its own target — one blended
account target is how conflicting bid logic sneaks back in.

#4.3 The bid equation and RPC bidding

Max CPC bid = Target ACoS × Price × CVR
  e.g. 30% × $29.99 × 0.11 = $0.99

RPC version (better once a target has ≥ ~50 clicks of its own data):
  RPC (revenue per click) = ad revenue ÷ clicks
  Max bid = RPC × Target ACoS
  e.g. RPC $3.40 × 30% = $1.02
RPC self-corrects for the keyword's real CVR and average order value.
New targets: start at 50–75% of Amazon's suggested bid (Elias), or 1.5×
suggested when deliberately forcing rank (AMZ 06 launch mode).

#4.4 Dayparting logic

Mechanics: all daily budgets refresh at midnight PT → a 3–6 hour "bid rush"
of expensive, low-converting clicks; budgets exhausted by early afternoon
miss the 6pm–midnight window where CPCs sag and CVR often peaks.
CVR can swing up to ~300% between best and worst hours.

  deviation_hour = (RPC_hour − RPC_avg) ÷ RPC_avg
  bid_hour = base_bid × (1 + 0.5–1.0 × deviation_hour)

Practical (native rules can only RAISE bids on schedule):
  set base bid for off-peak → schedule +15–25% raises in proven peak windows.
Gate: ≥ ~1,000 clicks of hourly data (SP hourly report via API/tools)
before dayparting anything. Below that you are fitting noise.

---

#5. 2025–26 BENCHMARKS — WHAT NORMAL LOOKS LIKE

Blended figures hide huge category variance — benchmark inside your category. Ranges triangulate Ad Badger, Pacvue, and SellerPlex 2025–26 data; bands, not gospel.

By category (CPC, 2025–26):

Planning rule: model next year's CPC at +8–15% — CPC inflation is structural (more sellers, more ad load, AI-assisted bidding everywhere). Your CVR and margin have to outrun it.

---

#6. THE OPTIMIZATION SOP — 45 MINUTES A WEEK

#6.1 The weekly routine (same day, same order, every week)

Monthly add-ons (60–90 min): SQP review (§8.3), SB/SD creative rotation, dayparting review, per-ASIN TACoS (kill or fix chronic bleeders).

#6.2 Bulk files and when tools take over

• Bulk operations (Seller Central → bulk spreadsheets) become mandatory around 50+ campaigns: download, filter by ACoS band, mass-edit bids, re-upload — the weekly routine drops from hours to minutes. Bradley Sutton's demo standard: 200 campaigns managed in ~10 minutes/week once rules are templated.
• Manual vs software threshold: manual + bulk files up to roughly $10–20K/month spend or ~50 campaigns; past that, rule-based or AI tooling pays for itself in reclaimed hours and faster negation cycles.

#6.3 The 2026 tool landscape (rough pricing — verify before signing)

Elias's warning: an algorithm optimizing ACoS without your COGS and TACoS context will happily strangle organic rank. Feed tools your margin data and run suggest-mode for the first 60 days.

---

#7. ADVANCED LAYER — AMC, DSP, AND THE RETENTION STACK

#7.1 Amazon Marketing Cloud (AMC)

Free clean-room analytics (SQL or templates; every major tool wraps it). Worth real attention from ~$50K+/month brands; CES 2025 extended purchase-signal lookback from 13 months to 5 years. Four operator use cases:

1. NTB% by campaign — which campaigns acquire new customers vs recycle existing buyers. Kills the "2% ACoS hero campaign" myth when it turns out to be 90% existing customers.
2. Path-to-purchase — new buyers in considered categories typically touch 3–5 ad exposures before converting; shows whether SB/SD/DSP exposure actually raises SP conversion (exposed-to-both cohorts routinely convert at multiples of single-channel cohorts).
3. Custom audiences — build segments standard targeting can't (e.g., "clicked twice, never bought, last 60 days"; "bought ≥2× but not in 90 days") and activate them in SD and DSP.
4. Incrementality checks — brand-term cannibalization and promo overlap analysis (§8.1).

#7.2 DSP — entry criteria (2026)

• Self-serve: Amazon dropped the imposed minimum at unBoxed 2025 — but practical floor is $10–15K/month so the algorithm has conversion signal. Mid-market brands typically ride an agency seat (BTR Media, Trivium, etc.) at 10–15%-of-spend fees.
• Managed service (Amazon-run): ~$50K minimum (varies by country); usually less hands-on than a good agency seat.
• Entry gate: full Sponsored-suite maturity first (TACoS stable ≤15%, SD retargeting profitable). DSP amplifies a working machine; it cannot fix one. Pacvue Q1 2025: DSP drove 36.5% NTB sales share for advertisers using it — its job is new demand and retention, not last-click ROAS.

#7.3 Retargeting economics (SD before DSP)

Retargeting profit test:
  Detail page CVR ≈ 10% → 90% of clickers leave without buying
  SD views-remarketing (30-day lookback) recaptures at CVR typically
  1.5–3× cold traffic, at CPCs often below head-term SP.
  Budget: 10–20% of total ad spend, scale only while ACoS ≤ efficiency target.
  Lookback by price point: <$25 impulse → 7–14 days; considered/premium → 30–90.
  Branded creative in SD ≈ +13.6% ROAS vs unbranded (Amazon data).

#7.4 Brand Metrics, Brand Tailored Promotions, Subscribe & Save

• Brand Metrics (free): awareness→consideration→purchase funnel vs category peers; watch branded-search growth as the upper-funnel scoreboard (also the Sponsored TV KPI).
• Brand Tailored Promotions (BTP): targeted discounts to behavioral segments — cart abandoners, brand followers, repeat and high-spend customers. Cart abandoners redeem at 15–25% (5–10× generic coupons); fewer than ~25% of eligible brands use BTP systematically. Coordinated ads+promos programs report 15–20% TACoS reduction while growing sales; track the blend as TPCS = (ad spend + promo cost) ÷ total revenue.
• Subscribe & Save × ads: on subscribable ASINs, bid on LTV, not first-order margin:

LTV-adjusted breakeven ACoS ≈ margin% × (1 + expected repeat orders)
  e.g. 40% margin, 2.5 expected reorders → can pay up to ~100%+ first-order
  ACoS and still profit on the subscription tail. This is how supplement
  brands survive $3–7 CPCs — and why you can't copy their bids without
  their reorder curve.

---

#8. DEFENSE & OFFENSE — BRAND TERMS, CONQUESTING, SHARE TRACKING

#8.1 Brand-term defense (when it's worth it)

• Economics: brand campaigns run 2–15% ACoS and should hold ≥95% impression share on your brand terms. Cost is usually 5–15% of total spend.
• Worth it when: competitors actively bid your brand (search your brand in incognito — if a rival ad sits above your organic result, you're leaking your cheapest conversions), or your brand terms carry real volume (SQP shows it).
• The incrementality caveat: if no one attacks your brand terms, defense spend is partly buying traffic you'd get free. Test: pause brand defense 2 weeks, watch total brand-term sales (SQP/AMC), not ad-attributed sales. Flat total = you were cannibalizing; keep only a token floor bid. Falling total = defense was real; restore it.

#8.2 Competitor conquesting math

Conquest CVR ≈ 0.4–0.7 × your category-keyword CVR (shoppers searching a
rival's brand mostly want the rival). Same bid equation, discounted CVR:
  Conquest max bid = Target ACoS × Price × (0.4–0.7 × base CVR)
  → conquest bids must run 30–60% BELOW your category-term bids for equal ACoS.
Run conquest only with a visible edge on the target ASIN: lower price,
review count within ~30% at better rating, or a real feature gap.
Judge conquest on NTB% (AMC) as much as ACoS — stolen customers compound.

#8.3 Category share tracking via SQP

Search Query Performance (Brand Analytics) is your market-share instrument: per query, your share of impressions → clicks → add-to-carts → purchases vs the whole market. Monthly ritual, top 20 queries: impression share growing but purchase share flat = CTR/CVR problem (AMZ 05, not more bids); purchase share falling while rank holds = a competitor is out-converting you — find out why before out-bidding them.

---

#9. FAILURE MODES — HOW OPERATORS DIE HERE

---

#MODULE SUMMARY — THE TEN COMMANDMENTS OF AMAZON PPC

1. TACoS is the north star; ACoS is a campaign dial. A falling TACoS with rising ACoS is winning; the reverse is dying with a pretty dashboard.
2. Know breakeven ACoS per ASIN before touching a bid — it equals pre-ad contribution margin, and every target ACoS is a multiple of it by lifecycle posture.
3. Structure by objective (Awareness · Rank · Efficiency · Branded · Conquesting · Auto): one job per campaign, single-keyword campaigns for heroes only, tight themes for the rest.
4. Run the harvest loop weekly: ≥2–3 orders at ≤ target ACoS → promote to exact + negative-exact the source; ~25–30 clicks with zero orders → negate. This loop is PPC management.
5. Bid = Target ACoS × Price × CVR (graduate to RPC × Target ACoS at ≥50 clicks); start new targets at 50–75% of suggested, 1.5–2× only when forcing rank.
6. Down-only dynamic bidding is the default; up-and-down only on proven profitable campaigns; isolate top-of-search with dedicated campaigns — it converts 2–3× rest-of-search.
7. Benchmark inside your category, in ranges: ~$1.10–1.20 blended CPC (supplements $2.50–7), 25–38% ACoS, 8–15% CVR, 0.4–0.6% CTR — and plan for +8–15% CPC inflation every year.
8. 45 minutes weekly, same order every time: search terms → negatives → bids by ACoS band → budgets → placements → rank/TACoS check. Bulk files past 50 campaigns; software past $10–20K/month — always fed your margin data.
9. Earn the funnel in order: SP efficient → SB/SD layered → AMC for NTB and path-to-purchase → DSP at $10–15K/month practical minimum → Sponsored TV only as a $250K+/month brand.
10. Defend brand terms at ≥95% impression share only if they're attacked (test incrementality before paying for your own traffic), conquest with bids discounted 30–60% for lower CVR, and read SQP monthly — purchase share, not impression share, is the scoreboard.