Amazon bidding strategy looks deceptively simple. You choose fixed bids, dynamic bids down only, or dynamic bids up and down. Maybe you add a Top of Search placement multiplier. Maybe you lower bids on weak keywords and push bids on winners. The interface makes it feel like a settings decision.
It is not. For a brand owner managing Amazon Ads without a full agency team, bidding strategy is a profit permission system. It decides how quickly Amazon is allowed to spend when the algorithm thinks a click might convert. That is useful when the product has margin, stock and conversion evidence. It is expensive when the product only has optimism and a nice-looking suggested bid.
The named mistake I see is letting Amazon’s conversion probability override your contribution margin reality. A campaign uses dynamic up and down on a €1.20 base bid, adds a 150% Top of Search multiplier, and suddenly a click can enter the auction with a theoretical ceiling of €6.00. Amazon may be right that the click is more likely to convert. That still does not mean the SKU can afford it after referral fees, FBA fees, VAT treatment, coupons, returns and the next replenishment cost.
My stance: do not choose an Amazon bidding strategy by campaign mood. Choose it by profit permission: the maximum CPC a SKU, keyword and campaign role can absorb today without breaking the commercial job you gave that campaign.
This guide is for brand owners managing marketplace advertising themselves, usually from around €1.5K monthly ad spend. FiveX helps here by connecting advertising data to SKU margin, stock cover, product profitability and automation rules, so bids are not judged by ACOS alone.
What current Amazon bidding advice gets right
The existing advice is useful. Amazon explains the mechanics clearly: down only can reduce bids when a sale looks less likely, while up and down can also raise bids. Amazon says a $1.00 bid can become $2.00 and positions the strategy for sales-focused, high-performing campaigns, deal ASINs and excess inventory.
Perpetua and BidX show how dynamic bidding interacts with placement adjustments. The important lesson is compounding. With a 900% placement multiplier and dynamic up and down, a €1.50 base bid could theoretically become €30.00 for Top of Search. Extreme, yes. Useful mental model, absolutely.
Teikametrics explains the bid formula many advertisers use: average order value multiplied by conversion rate multiplied by target ACOS. SellerMetrics adds a practical launch view: new products often belong on dynamic down only because Amazon has limited sales history.
Reddit threads add the part vendor blogs sometimes smooth over. Sellers complain that up and down often raises CPC and ACOS, suggested bids can feel inflated, and low down-only bids may receive almost no impressions. That is the operator reality: every setting has a cost.
What most advice misses is the layer between the auction and finance: a daily rule for what Amazon is allowed to do with this SKU, in this campaign role, with this margin, this stock cover and this search intent.
The missing layer: profit permission before bid permission
Amazon’s bidding strategies optimise inside Amazon’s signal set. That signal set is powerful, but it does not know your full business case: an 11% landed-cost increase, twelve days of stock left, a finance cap until payout clears, or a coupon that changed break-even ACOS yesterday.
That is why bidding strategy should start with a number Amazon does not give you: break-even CPC by SKU and campaign role.
Break-even CPC = contribution margin per order × expected conversion rate
If a product earns €8.40 contribution margin after marketplace fees, fulfilment, product cost, expected returns and discounts, and a campaign converts 10% of clicks into orders, the break-even CPC is €0.84. At €0.84, the next click is expected to consume the entire contribution margin. If you want the campaign to keep 30% contribution after ads, the permitted CPC is €0.59. That is the number your advertising software should compare against Amazon’s bid setting.
ACOS is still useful, but blunt. A 25% ACOS can be wonderful on a product with 45% contribution margin and painful on one with 18%. A 40% ACOS can be acceptable in a launch campaign if stock, reviews and learning budget support it. Bidding only makes sense after role clarity.
The four campaign roles that should decide bidding strategy
I separate Amazon Sponsored Products campaigns into four practical roles, because bidding without role clarity quickly becomes messy.
1. Discovery campaigns
Discovery campaigns exist to learn through auto targeting, broad match, category targets or competitor ASINs. Their job is search-term evidence without accidentally funding a small bonfire.
Default rule: start with dynamic down only or fixed bids, conservative daily budget and a CPC cap below the SKU’s full profit ceiling. Use FiveX to monitor search terms, spend, stock and margin together, then promote converting terms only after they pass your evidence threshold.
2. Conversion campaigns
Conversion campaigns target proven exact keywords or product targets. Dynamic up and down can make sense here, but only when conversion history is real and the maximum possible CPC still fits profit permission.
Default rule: allow dynamic up and down only when the last 14 to 30 days show stable conversion, enough clicks, acceptable returns and no stock constraint. Five orders from seven clicks is lovely, but not enough to rewrite policy.
3. Defence campaigns
Defence campaigns protect branded terms, own ASINs and valuable product pages. They often look efficient because the shopper already knows the brand. That does not mean they deserve unlimited bids; sometimes a lower fixed bid with placement control is better.
Default rule: keep defence campaigns on fixed or down only unless competitor pressure is visible and stock is healthy. In FiveX, share of voice, organic ranking, ad spend and product margin should sit in one view. You are buying risk reduction, not vanity ownership.
4. Clearance campaigns
Clearance campaigns exist when stock economics change: aging inventory, storage fees, seasonal deadlines or a packaging refresh. In that case, a lower contribution target may be rational.
Default rule: dynamic up and down can be allowed with a clear end date, capped total budget and a lower profit target. The mistake is leaving those settings active after the clearance reason disappears. Automation should expire with the commercial context.
Named scenario 1: Luna Pet Supplies and the €6 Top of Search trap
Luna Pet Supplies sells a premium dog grooming brush on Amazon.de for €34.95. After referral fees, FBA, landed cost, expected returns and a small coupon, the contribution margin before ads is €9.10. The exact keyword “dog grooming brush long hair” converts at 12% over the last 30 days. Break-even CPC is therefore €1.09. If Luna wants to keep 25% contribution after ads, the permitted CPC is about €0.82.
The campaign has a €1.20 base bid, dynamic up and down and a 150% Top of Search multiplier. In plain English: Luna has told Amazon it may bid far above the CPC the SKU can actually afford. Even if the average CPC lands at €1.35 instead of the theoretical maximum, the campaign needs a stronger conversion rate than history supports.
The better setup is role-based permission. Keep the exact keyword in a conversion campaign. Set the base bid around €0.70 to €0.80. Allow a modest Top of Search multiplier only if that placement converts better than rest of search. Set a FiveX alert if CPC rises above €0.90 for three days or stock cover drops below 21 days. The campaign can still compete; it just cannot spend as if margin is infinite.
Named scenario 2: NordicHome and the launch campaign that needs learning, not bravery
NordicHome launches a bamboo drawer organiser on Amazon.nl at €24.95. Contribution margin before ads is €6.20. There are four reviews, little sales history, and a €1,800 first-month ad budget. Amazon suggests €0.95 to €1.80 bids for several generic terms.
If NordicHome chooses dynamic up and down immediately because it wants velocity, it is letting Amazon buy confidence the listing has not earned yet. At a 7% early conversion rate, break-even CPC is €0.43. A €1.10 CPC would require a conversion rate near 18% just to break even before considering learning waste.
The better strategy is boring in the nicest possible way. Use dynamic down only on discovery campaigns. Cap generic broad terms at €0.35 to €0.45. Give the auto campaign €20 per day. Move any search term with at least two orders, 15%+ conversion and CPC below €0.55 into an exact campaign. FiveX can turn that into a weekly search-term handover.
Named scenario 3: FitFuel and the clearance exception
FitFuel sells protein bars in multipacks. A vanilla flavour has 1,200 units left, a best-before date in ten weeks and a normal contribution margin of €4.80 per order. The product usually cannot tolerate CPCs above €0.55, but holding the stock creates a different cost: discounting later, storage pressure and cash tied up in aging inventory.
Here, aggressive bidding can be rational. FitFuel creates a clearance campaign with dynamic up and down, a €60 daily budget, a four-week end date and a lower contribution target. It excludes generic research terms and accepts 38% ACOS instead of the usual 24% target.
The important part is the expiry rule. When stock drops below 250 units or the end date arrives, FiveX should flag the campaign and return the SKU to normal profit permissions. Clearance logic is a temporary commercial exception, not a new default.
How to choose the right bidding strategy
Use this operating sequence before changing campaign settings.
Step 1: Calculate the SKU’s current contribution margin
Include selling price, referral fee, fulfilment, product cost, shipping, expected returns, discounts, tax treatment and marketplace-specific costs. Do not use last quarter’s margin if fees or coupons changed. FiveX’s product profitability view keeps this live for ad decisions.
Step 2: Estimate conversion rate by campaign role
Do not use the account average. A branded exact campaign, a generic broad campaign and a competitor ASIN target are different animals. Use the last 14 to 30 days when volume is stable.
Step 3: Turn margin and conversion into CPC permission
Contribution margin multiplied by conversion rate gives break-even CPC. Then apply the role: discovery might receive 50% to 70% of it, conversion 70% to 95%, defence depends on competitor pressure, and clearance may temporarily exceed normal limits.
Step 4: Calculate the maximum possible bid, not just the base bid
This is where many teams get caught. Dynamic up and down can raise your base bid; placement multipliers can compound it. Before activating the setting, ask: “What is the highest bid Amazon could theoretically enter with?” If finance twitches, lower the base bid or multiplier.
Step 5: Add stock and Buy Box guardrails
A profitable CPC is not profitable if the SKU is about to stock out. FiveX can connect Amazon Ads performance to inventory cover and Buy Box signals, so automation does not push a product that cannot safely receive demand.
Step 6: Review search terms before changing strategy
High ACOS does not always mean the bidding strategy is wrong. Sometimes the campaign is buying the wrong terms. Before switching from down only to up and down, check whether broad discovery and exact conversion live together. If so, split them first.
When fixed bids are the right choice
Fixed bids are not old-fashioned. They are useful when you need predictability. Use them when the product has thin margin, the campaign is defending branded traffic, the data sample is unstable, or you are testing whether Amazon’s dynamic decisions are actually adding value.
A good fixed-bid test is simple: take one proven keyword, clone the structure, keep budget controlled, and compare fixed versus dynamic down only over two weekly cycles. Judge contribution after ads, not just ACOS.
When dynamic down only is the right choice
Dynamic down only is the safest default for learning, unstable data and early-stage launches. Amazon can reduce bids when conversion probability looks weak, without permission to push above your base bid.
The trade-off is impression starvation. If bids are too low, the campaign may learn nothing. Raise bids gradually when relevant targets receive no impressions. Lower them when clicks arrive but intent is wrong. The setting protects the downside; it does not replace judgement.
When dynamic up and down is the right choice
Dynamic up and down belongs where evidence already exists. Proven exact keywords, strong product targets, deal periods, excess stock and high-performing campaigns can all justify it. The key is not whether the strategy can produce more orders. It often can. The key is whether the extra orders still respect the SKU’s margin and operational context.
Use it with three controls: a base bid below permitted CPC, placement multipliers that avoid absurd theoretical bids, and alerts when CPC, ACOS, stock cover or conversion rate breaks the permission band. That is FiveX-style self-service automation: not “set and forget”, but “set, monitor and intervene only when the commercial rule breaks”.
The weekly bidding review I would actually run
If you manage your own Amazon Ads, do not review every keyword from scratch every morning. That is how people become very busy and only slightly wiser. Run a weekly bid permission review instead.
First, list campaigns by role. Second, pull SKU margin and stock cover. Third, flag campaigns where actual CPC exceeds permitted CPC, ACOS exceeds role target, stock cover is low, conversion dropped, or maximum possible bid is more than 2.5 times permitted CPC. Fourth, act: reduce bid, split campaign, switch strategy, cap placement, pause target, or keep learning.
FiveX can automate much of this review by bringing ad spend, revenue, ACOS, ROAS, SKU margin, inventory, search-term performance and alerts together. The point is to surface the few bid decisions that actually deserve human attention.
Final thought: Amazon can optimise the auction, but you own the business case
Amazon bidding strategy is not a personality test. You are not “a dynamic up and down brand” or “a fixed bid brand”. You are a business deciding which clicks deserve money under today’s margin, stock and campaign conditions.
Let Amazon’s algorithm help where it has evidence. Let fixed bids protect decisions where you need control. Let dynamic down only gather learning without giving away the wallet. Let dynamic up and down scale proven demand when the SKU can afford it. And let your advertising software sit above all of it with profit permission rules that Amazon cannot infer from the auction alone.
The goal is not the cleverest bid setting. The goal is a marketplace ad system where every extra euro has a job, a ceiling and a reason to exist. Much less dramatic than chasing the top slot. Much better for profit.