Amazon Advertising history is usually told as a product timeline. Sponsored Products became easier to manage. Amazon unified AMG, AMS and AAP under the Amazon Advertising name in 2018. Sponsored Display arrived in 2019. DSP became more useful. Amazon Marketing Cloud gave larger advertisers cleaner measurement. Ad revenue climbed from a marketplace side business into a $56.2 billion line item in 2024.
That history is useful. It is also incomplete for operators.
If you manage marketplace advertising for a brand spending €5K, €25K or €250K per month, the practical lesson is not “Amazon keeps launching more ad formats”. You already know that. The lesson is sharper: every time Amazon adds reach, targeting or measurement, the account needs a new profit control. The brands that win are not the ones that adopt every format fastest. They are the ones that decide which SKU is allowed to use which format, at what margin, with what evidence, and when the experiment must stop.
The named mistake is timeline copying. A team reads the latest Amazon Ads evolution story and turns it into a roadmap: Sponsored Products first, Sponsored Brands next, Sponsored Display after that, DSP when we look mature, AMC when finance asks difficult questions. Nice deck. Dangerous operating model. Amazon’s history is not a maturity checklist. It is a series of permission changes.
This guide takes the 2017-to-now story and translates it into an operating model for Amazon, bol and MediaMarkt advertisers in the Netherlands and Belgium. The goal is simple: use the history to build better ad decisions today, not to admire the timeline.
What the research says — and what most articles miss
Podean’s history piece is helpful because it documents the big milestones. It notes how Amazon Advertising topped $10.1 billion in 2018 ad revenue, how personalised display ads and video ads expanded the toolbox in 2017, and how the 2018 rebrand retired Amazon Media Group, Amazon Marketing Services and Amazon Advertising Platform in favour of one Amazon Advertising brand. That matters because the rebrand was not cosmetic. It signalled that retail media was becoming a serious, unified commercial engine.
BidX adds the revenue lens. Its coverage of Amazon’s ad business highlights that Amazon disclosed advertising services more clearly in 2021 and that ad revenue had grown almost tenfold from 2016 to 2021. BidX’s 2024 performance analysis is also useful: across its platform data, ad spend increased 13.3%, ad sales grew 19.5%, ROAS improved from 5.0 to 5.3, ACOS dropped to 18.9%, and CPCs rose 9.2% to $0.61. Sponsored Products still carried 82.3% of spend, while DSP grew strongly as sellers moved further up the funnel.
Sponsored Display guides from BidX, Nozzle and others explain the format well. Sponsored Display began as a 2019 beta, replaced Product Display Ads, expanded across marketplaces, and made self-service audience and product targeting easier on and off Amazon. AMC articles explain the next measurement step: privacy-safe clean-room analysis, longer lookback windows, and more serious path-to-purchase questions.
Reddit and seller-forum search results show the less polished side. Sellers do not mainly ask whether Amazon Ads is historically important. They ask why PPC feels more expensive, how long to tolerate high ACOS, whether PPC is a necessary evil, and whether agencies understand SKU-level break-even points. That is the real operating pressure.
Here is the gap: most history articles explain what Amazon launched. Most tool articles explain what each feature does. Very few explain which profit control should have been added at each stage. That is the angle FiveX can own.
The real timeline: from buying clicks to buying permission
Before 2017, many marketplace teams treated Amazon PPC like paid search inside a shop. Find keywords, set bids, harvest converting terms, add negatives, repeat. That still matters. But from 2017 onward, Amazon Ads became something broader: display, video, audiences, retail media, DSP, AMC, streaming inventory, brand-building and closed-loop measurement.
Each step changed the operator’s job.
- Sponsored Products scale made keyword and product targeting a daily sales lever. The required control: SKU-level break-even ACOS, not one account-wide target.
- The 2018 unification into Amazon Advertising made ads feel like a strategic retail-media platform. The required control: campaign roles, so defence, discovery, conquesting and launch spend do not share the same KPI.
- Sponsored Display in 2019 made remarketing and product-page targeting more accessible. The required control: audience permission, because not every visitor is worth chasing after the margin stack changes.
- DSP growth moved spend beyond last-click product ads. The required control: incrementality rules, not just attributed sales.
- AMC and cleaner measurement made path analysis possible. The required control: decision rights, because more data can create more debate unless the team agrees what evidence is enough to act.
That is why a marketplace advertising service cannot simply “manage Amazon Ads”. It has to manage the permission system around Amazon Ads. FiveX connects advertising data with product profitability, inventory insights, repricing context and marketplace analytics so the ad decision is not trapped inside the ad console. Lovely charts are not enough. The bid needs to know whether the SKU can afford the click.
Control 1: Sponsored Products need margin permission, not nostalgia
Sponsored Products still does the heavy lifting in most Amazon accounts. That is exactly why it is dangerous. Mature formats become invisible. Teams assume the structure is “fine” because it has existed for years.
Imagine NorthPeak Home, a fictional Dutch kitchenware brand spending €7,200 per month on Amazon.de Sponsored Products. Its hero pan sells for €39.95. Referral and fulfilment fees take €10.80. Landed cost is €15.40. Returns and damage allowance add €1.60. A seven-day coupon removes €4.00. Contribution before ads is €8.15.
The campaign’s average CPC is €0.74 and conversion rate is 6.1%. Paid traffic therefore costs about €12.13 per order. In-platform ACOS looks acceptable at 28% because revenue is healthy. The SKU-level profit view says the promoted order loses roughly €3.98 before overhead during the coupon.
The wrong response is to pause all discovery and declare PPC too expensive. The better response is a margin permission rule: keep brand defence live, cap discovery budget at €45 per day, lower bids 18% on generic non-branded terms, and automatically re-open the test 48 hours after the coupon ends if stock cover remains above 21 days.
This is where FiveX’s product profitability and advertising automation hooks matter. The operator should not rebuild this calculation in a spreadsheet every Monday. The account should see the SKU margin, promotion state, ad cost and stock context together, then route the action with a clear reason.
Control 2: Sponsored Display needs audience permission
Sponsored Display changed the mental model. The question stopped being only “which search term converts?” and became “which shopper deserves another impression?” That is a different commercial decision.
Consider LumaPet, a fictional pet accessories brand selling on Amazon.nl and bol. It spends €5,600 per month across marketplace ads. A memory-foam dog bed has a selling price of €54.95, contribution before ads of €16.20 and a 30-day return rate that recently fell from 17% to 8.5% after packaging was improved. On Amazon, Sponsored Display retargeting visitors from the last 14 days produces a 22% ACOS. On bol Sponsored Products, the same SKU has a higher CPC but better net contribution because fulfilment cost is lower.
A format-led team would scale the Amazon audience because the ACOS is prettier. An operator-led team asks three questions first: are these shoppers incremental, is the SKU still profitable after returns, and is Amazon the best marketplace for the next euro?
The decision might be: increase Amazon Sponsored Display retargeting from €40 to €70 per day only for product viewers who did not buy, keep bol Sponsored Products at €60 per day for the Dutch “orthopaedic dog bed” query cluster, and block retargeting audiences if return rate climbs above 11% again.
That is not a generic retargeting tip. It is audience permission. FiveX helps here by putting advertising performance next to returns, SKU profitability and channel comparison. The team can decide whether an audience is worth buying, not merely whether Amazon allows targeting it.
Control 3: DSP and upper-funnel spend need evidence windows
DSP is where many marketplace teams become either too conservative or too impressed with themselves. Too conservative, and they never test upper-funnel media because last-click ROAS looks safer. Too impressed, and they fund awareness with money that should have stayed in profitable product ads.
The trade-off is not “DSP good” or “DSP bad”. The trade-off is evidence.
VoltEdge, a fictional electronics accessory brand, spends €8,400 per month across Amazon and MediaMarkt retail media. It wants to push a USB-C docking station before the back-to-office peak. On Amazon, Sponsored Products converts at 7.4% with a €0.68 CPC. On MediaMarkt Sponsored Product Ads, CPC is €0.52 but conversion is only 3.9%. The MediaMarkt audience looks weaker until the team notices that basket value is higher and price position improved from 4% above market to 3% below after repricing.
An Amazon-history deck might say: “Now test upper funnel.” A profit-control operating model says: “Define the evidence window before you spend.” For VoltEdge, that could mean a €1,200 DSP or retail-media prospecting test with a 21-day read, a maximum blended TACoS lift of 1.8 percentage points, and a requirement that branded search volume or detail-page sessions rise by at least 12% in the test segment. If those signals do not appear, the budget rolls back to Sponsored Products and MediaMarkt bottom-funnel placements.
FiveX’s repricing context, inventory insights and AI recommendations support that decision. If price position improved, stock cover is 36 days and margin is healthy, the test deserves oxygen. If stock falls to 9 days, the same campaign should not keep buying demand just because the audience model looks clever.
Control 4: AMC-style measurement needs decision rights
AMC and clean-room measurement are powerful because they let teams ask better questions: How many touchpoints happen before purchase? Which paths include DSP and Sponsored Products? Do returning buyers behave differently from new-to-brand shoppers? How long should the lookback window be?
But more measurement can also create a meeting swamp. Everyone finds a different slice that supports their preference. The PPC lead wants to keep the keyword. The brand lead wants more upper funnel. Finance wants contribution margin proof. Sales wants volume. Operations wants fewer stockouts. All are reasonable. None should own the decision alone.
So the control is decision rights. Before running advanced analysis, define who can approve each move:
- Bid changes within existing margin rules: PPC owner.
- Budget shifts between Amazon and bol: marketplace lead plus finance threshold.
- DSP or Sponsored Display tests above €1,000: growth lead with a defined evidence window.
- Pauses caused by stock, Buy Box or delivery-promise risk: operations veto.
- Target ACOS changes after cost updates: finance-approved margin version.
This is the unglamorous part of modern Amazon Advertising. The platform has become sophisticated enough that “the ad specialist decides” is no longer always safe. FiveX turns those cross-functional signals into a shared operating view, so ad recommendations can be accepted, challenged or paused with the same source of truth.
How to use Amazon Advertising history in your account this week
Do not turn the timeline into a shopping list. Turn it into an audit.
- Map every campaign to a role. Defence, discovery, launch, conquesting, retargeting, clearance and upper funnel should not be judged by one blended ACOS.
- Add SKU-level break-even points. If the account only has category-level ACOS targets, Sponsored Products is probably hiding margin leaks.
- Give each audience a permission rule. Sponsored Display and retargeting should know when returns, stock, price position or margin make the audience temporarily unattractive.
- Pre-define evidence windows for bigger tests. DSP, video and retail-media pilots need a budget cap, read period, primary signal and rollback rule before launch.
- Separate platform history from business maturity. You do not “deserve” DSP because Amazon launched it years ago. You deserve it when your unit economics, inventory and measurement discipline can support it.
That last point is the stance. Amazon Advertising has evolved from a click-buying tool into a retail media ecosystem. Good. But the more advanced the ecosystem becomes, the more expensive lazy governance gets.
The FiveX view: history is only useful if it changes the next decision
The best lesson from 2017 to now is not that Amazon became an advertising giant. It is that marketplace advertising moved closer to the P&L. Ads now touch margin, inventory, price position, customer acquisition, organic rank, marketplace mix and cash timing. Managing that from separate exports is possible at small scale. At €5K+ per month, it becomes fragile.
FiveX is built for that operating reality. We connect marketplace analytics, advertising automation, SKU profitability, inventory insights, repricing and AI recommendations so teams can decide where the next euro should go across Amazon, bol and MediaMarkt.
Use Amazon Advertising history as a warning, not a trophy cabinet. Every new format gives you more reach. Every new measurement layer gives you more explanation. Neither guarantees profit. Profit comes from the controls around the format: margin permission, audience permission, evidence windows and decision rights.
That is the operator lesson. The platform evolved. Your governance has to evolve with it.