LOWER ACOS, HIGHER PROFIT

Amazon PPC (Pay-Per-Click) is an advertising system in which sellers bid on keywords or competitor ASINs and pay only when the ad is clicked, not when it is shown. In every account we manage at Mentoreis, the core of the work rests on a single mechanism: Amazon runs a second-price auction. That means the seller who wins the auction does not pay their full bid, but one increment above the second-highest bid. This is why aggressive bidding alone does not pay off; it only pushes competitors' payment thresholds higher.

The second variable that decides the auction is relevance. For every candidate ad, Amazon evaluates the estimated click-through rate, the historical conversion performance on that search term, and the semantic fit between the listing and the query. Between two sellers competing with the same $1.20 bid, the one with the higher conversion rate ranks higher and pays less. In practice this means one thing: a poor listing puts a hidden tax on your PPC budget. Advertising management and listing optimization are two sides of the same system.

The third critical detail is the attribution window. Amazon credits a purchase made in the days following a click back to the ad that was clicked. The most commonly confused detail here is this: for Sponsored Products, the click attribution window for seller accounts (Seller Central) is 7 days; the 14-day figure you often hear applies to vendor (1P) accounts. For Sponsored Brands and Sponsored Display the window is 14 days. So a customer clicks the ad, does not buy that day, then comes back through organic search on the fifth day and buys; the sale still shows up in the advertising report. This is why changing bids based on 48 hours of data is misleading, the data has not matured yet. At Mentoreis the standard optimization cycle is 7 days and the permanent decision cycle is 14 days.

Finally, PPC is not only a sales channel. Every sale that comes through advertising feeds Amazon's ranking signals: search-term-level conversion, sales velocity and customer engagement. As conversions accumulate on a keyword through advertising, organic rank on that same keyword rises. A properly structured PPC account reduces its own need for advertising over time; a poorly structured one demands a bigger budget every month.

ACOS (Advertising Cost of Sales) is the ratio of ad spend to ad-attributed revenue. If you spent $1,000 and generated $5,000 in ad revenue, your ACOS is 20%. TACOS (Total ACOS) is the ratio of ad spend to total revenue, meaning the sum of advertising and organic sales. The single most important indicator of a brand's real health is the gap between these two.

There is no target ACOS figure that fits everyone. Target ACOS is derived from margin. A concrete example: a $29.99 product, $6.00 product cost, $5.69 FBA fulfillment, 15% referral fee which is $4.50, plus $0.80 for storage and returns. Total cost is $16.99 and unit contribution margin is $13.00. Break even ACOS in this case is 43.3%. For this seller a 35% ACOS is not a loss, a 55% ACOS is. The first question to ask when choosing an agency is this: on what basis do you set the target ACOS?

TACOS is read through its trend over time. In a healthy account ACOS stays flat while TACOS falls, because the organic share of sales grows. If TACOS is above 20% and rising, the brand is failing to take root organically; the moment advertising stops, sales stop too. For a mature brand, a TACOS band of 5% to 10% is considered ideal. During launch, a TACOS of 25% to 40% is temporarily normal, but it should come down within 90 days.

Setting bids by instinct is the most expensive habit there is. The formula we use is straightforward: Target CPC = Selling Price x Conversion Rate x Target ACOS. On a $29.99 product converting at 12%, if you are targeting a 30% ACOS, the reasonable maximum bid is 29.99 x 0.12 x 0.30 = $1.08. Every cent above this threshold is mathematically a loss. In the large majority of the accounts we audit, bids sit 40% to 90% above this calculated ceiling.

Campaign architecture is the backbone of an Amazon advertising account and the most expensive component to fix after the fact. Mentoreis builds every account on a three-layer structure: discovery, harvest, defense. Without this separation it is impossible to tell which spend is collecting data and which is producing profit.

The discovery layer consists of automatic campaigns (close match, loose match, substitutes and complements in separate ad groups), broad and phrase match, and category targeting. The purpose here is not sales but data: which real search terms convert. The discovery layer takes 15% to 20% of the total budget and runs with low bids and wide reach.

The harvest layer is the profit center of the account. Search terms that convert in discovery are moved into exact match campaigns, grouped into tightly themed ad groups, and bid individually according to their own conversion data. This layer takes 55% to 65% of the budget. The critical rule: every harvested keyword is blocked as a negative exact in the discovery campaign it came from. Otherwise your own campaigns bid against each other in the auction and you push your own CPC up.

The defense layer covers your brand keywords, cross-selling on your own detail page, competitor ASIN targeting and Sponsored Display remarketing. Brand defense campaigns typically run in the 8% to 15% ACOS band and stop competitors from stealing traffic off your product page. On format distribution, our practical weightings are as follows: Sponsored Products at 70-75% of budget, Sponsored Brands at 15-20% (especially SB Video, which produces 2-3 times the CTR of static formats), Sponsored Display at 5-10%. On the Amazon DSP side there is a widespread piece of misinformation: Amazon removed the minimum spend requirement for self-service DSP at the end of 2025, so there is no longer a platform-imposed floor. The $50,000 figure you often hear is the ad spend threshold for Amazon's managed service, not a revenue figure. We recommend DSP for brands that have a defined audience strategy and a solid Sponsored layer; the practical starting threshold is $10,000 to $15,000 in monthly DSP spend. For brands below that budget, Streaming TV ads, which carry no minimum spend, are a more realistic upper-funnel option.

Phase 0, Audit (Day 1-7). We read the account's 60-day search term report, placement breakdown, campaign structure and Brand Analytics Search Query Performance data together. In the same week we build the profitability model: we calculate break even ACOS from product cost, FBA fees, referral, storage, return rate and coupon expenses. In the accounts we take over, we typically find that 18% to 32% of the ad budget went to search terms that produced not a single conversion in the last 60 days.

Phase 1, Restructuring (Day 8-30). The three-layer architecture is built, the negative keyword layer is written, bids are reset using the formula and placement multipliers are set. Top of Search placement produces the highest conversion in most categories; a 30% to 50% modifier here generally grows revenue without raising ACOS. Throughout this phase ACOS fluctuates temporarily, which is expected and is stated in advance in the report.

Phase 2, Optimization (Day 31-60). Bid adjustments, search term harvesting and dead keyword pruning are carried out in seven-day cycles. Our pruning rule is concrete: a search term is negated if it has taken twice the inverse of the product's conversion rate in clicks and produced no sales. On a product converting at 10% that threshold is 20 clicks; on a product converting at 5% it is 40 clicks. This single rule cuts 15% to 25% of wasted spend in the first month on a typical account.

Phase 3, Scaling (Day 61 onward). Budget and bid increases on profitable keywords, hour and day level dayparting, new keyword and competitor ASIN targets, Sponsored Brands Video and the Sponsored Display remarketing layer all come into play. During Prime Day and Q4, CPCs inflate by 30% to 60% depending on category; the budget and bid plan for these periods is set at least three weeks in advance. Reporting is not campaign based but ASIN and search term based: a short weekly summary and a detailed monthly profitability report.

The first mistake is putting everything into one campaign. When automatic targeting, broad keywords and exact keywords sit in the same campaign, you cannot measure which spend produced what, and the budget always flows to the term that takes the most clicks and converts the least. The second mistake is never building a negative keyword layer. In our audits, accounts with zero negative keywords are still the majority.

The third mistake is daily optimization. Changing bids every morning before the attribution window has closed constantly resets the algorithm's learning period and destabilizes CPC; on Sponsored Products that window is 7 days for seller accounts. The fourth mistake is mistaking ACOS for profit. ACOS is a cost ratio, not profit; any ACOS decision made without accounting for coupon, return and storage expenses is made on incomplete data. The fifth mistake is trying to rescue a low-converting listing with advertising. If the conversion rate is below 6%, the problem is not in the advertising but in the image, the price or the reviews; raising the budget only accelerates the loss.

The sixth mistake is leaving brand keywords undefended. If you do not advertise on your own brand name, your competitors buy that space and you hand them your cheapest conversion. The seventh mistake is thinking about inventory planning separately from advertising. If you run out of stock while scaling a campaign, the listing loses its organic rank, the campaign history is damaged and recovery typically takes 3 to 6 weeks. At Mentoreis the advertising plan is always built together with inventory capacity.

The honest timeline in Amazon PPC management is this: the first 30 days are structure and cleanup, days 30-60 are stabilization, days 60-90 bring measurable improvement. An agency promising an ACOS drop within the first two weeks is either cutting advertising and cutting revenue with it, or reading the data before it has matured.

Typically the most visible gain in the first 30 days is cutting waste. By pruning search terms that produce no conversions and pulling bid ceilings back to the formula, you start producing the same revenue with less spend. By day 60, average CPC falls because the share of harvest campaigns has grown, and ACOS moves closer to the margin target. On day 90 the real indicator to watch is not ACOS but TACOS: because organic rank rises on the keywords fed by advertising, total revenue grows while the advertising share shrinks.

The five core indicators we measure and report are these: ACOS, TACOS, average CPC, ASIN-level conversion rate and organic rank on target keywords. On top of these we track market share metrics through Search Query Performance: how much of the total impressions, add-to-carts and purchases on a search term you capture. This report is the only official Amazon data that shows your real position against your competitors.

We say plainly that results cannot be guaranteed. Advertising management will not rescue a bad product, an uncompetitive price or a review profile averaging 3.5 stars. But if the product is right, the price is right and the listing is solid, PPC is the fastest lever for moving that product up the category rankings.

Mentoreis does not run PPC management tied to a single piece of software. When making decisions, our priority is always Amazon's own first-party data: Seller Central search term reports, placement reports, unit session percentage from Business Reports, and Search Query Performance within Brand Analytics. That last report is the only official source that tells you how many impressions, clicks, add-to-carts and purchases a search term generated, and how much of that belongs to you.

On the operational side we work through the Amazon Ads API and bulk operations files. Changing bids on thousands of targets one by one through the interface is both slow and error-prone; changes made with bulk files can be versioned and rolled back. For decisions that require hourly data we use Amazon Marketing Stream: we make dayparting decisions, meaning at which hours of the day bids are raised or lowered, with hourly conversion data rather than guesswork.

For keyword research and competitor analysis we use Helium 10 (Cerebro reverse ASIN and Magnet), SellerSprite and Data Dive. These tools provide estimated volume; they make no claim to precision. That is why no keyword is ever budgeted based solely on a tool's volume estimate; the decision is always validated against your own account's search term report.

In reporting, raw Amazon data is combined with the profitability model. The advertising console shows you ACOS but not net profit, because it does not know your FBA fees, return rate, storage or aged inventory costs. Mentoreis reports include post-advertising unit profit for every ASIN; decisions are made on that line.