HOW AMAZON ACTUALLY WORKS
What the algorithm is really testing — and how operators should respond
A Retail Readiness Field Guide
Steven Bickers | retail.draymoorventures.com
"Retail gave us visibility. Amazon gave us velocity."
1. Amazon Is Not a Retailer
Most brands approach Amazon the way they approach a retail buyer — build a relationship, tell the story, get approved, then let momentum carry. That framework is wrong, and it's expensive to learn at scale.
Amazon doesn't have a buyer in the traditional sense. There are category managers, and they matter, but they are not the decision-maker in the way a Target or Walmart merchant is. On Amazon, the system decides. The humans interpret what the data is already saying.
This is the foundational distinction. You are not pitching a person. You are performing for an algorithm that has no memory, no goodwill, and no patience for narrative.
"You don't win Amazon. You earn it — every day."
2. What the Algorithm Is Actually Testing
Amazon's algorithm continuously evaluates six variables. These aren't hidden. But brands consistently underestimate how unforgiving the system is when any of them slip.
Conversion rate. If customers land on your listing and don't buy, Amazon interprets that as product failure, not marketing failure. The algorithm doesn't know you had a bad photo or a price mismatch. It knows customers didn't convert. Visibility drops.
Velocity. Units sold per day relative to category peers. A new product gets a temporary visibility window — essentially a test. If velocity doesn't hit within that window, organic rank falls quickly. Brands often mistake the launch window for earned position. It isn't.
Returns. A high return rate signals product misrepresentation or quality failure. Amazon protects its customer experience above everything else. Products with rising return rates face suppressed visibility and, at certain thresholds, listing suspension without warning.
Reviews. Volume and recency both matter. An older product with 2,000 reviews and no recent activity trends differently than a product with 800 reviews and consistent recent volume. Amazon weights recency more than most brands realize.
Price competitiveness. Amazon's algorithm monitors your price against comparable listings and against your own history. If you raise prices — even for legitimate margin reasons — and conversion drops, rank follows. Price is not just a margin lever. It's a behavioral signal.
Inventory reliability. Stockouts are penalized in rank and are difficult to recover from. Consistent in-stock rates are a competitive advantage on Amazon in a way they are not on any physical channel. The algorithm rewards predictability.
"On Amazon, you're not approved once. You're re-approved every day."
3. Early Traction Is Often Misleading
This is the failure pattern I've seen most often, including in categories I've operated in. A brand launches on Amazon, gets early velocity from launch promotions or a small influencer push, builds reviews, and watches rank climb. Internally, that reads as validation.
Amazon hasn't made a judgment yet. It's collecting signal.
The questions the algorithm is answering during the first 60–90 days: Do customers repurchase? Do returns stay low? Does pricing hold under competitive pressure? Can supply remain consistent as velocity increases?
Brands that mistake the launch window for earned position make downstream resource allocation errors. They invest in expanding the Amazon catalog before the core SKU has proven durability. They reduce promotional support because organic rank looks stable. Then the rank slips and they can't figure out why.
The algorithm didn't change. The temporary inputs that created the early rank advantage ran out.
4. Category Shifts Are Ruthless
Amazon reacts to category shifts faster than any physical retailer I've worked with. When customer expectations change — a new entrant redefines what the product should cost, or a feature becomes table stakes — Amazon doesn't wait for legacy products to adapt.
Search behavior changes. Conversion patterns change. Recommendations change. There is no call from a category manager to discuss what happened. There is no reset meeting. The system simply routes around products that no longer match the new expectation set.
The implication for brands: you need to know where your category is going before you're executing against where it's been. The brands that get caught by category shifts on Amazon were usually watching the wrong indicators — their own velocity, their own reviews — rather than what the algorithm was starting to reward at the category level.
Price, feature set, and format all reset when a category redefines itself. Being the second-best product in the old category is not a defensible position.
5. Price Is a Signal, Not a Strategy
Operators often treat Amazon pricing as a margin management exercise. Raise the price when costs go up. Run a promotion to clear excess inventory. Layer in coupons to boost conversion for a ranking campaign.
That's not wrong, but it misses how the algorithm interprets pricing behavior. Price is a behavioral input that affects conversion, which affects rank, which affects visibility, which affects velocity. Changes compound quickly in both directions.
When price drifts above the competitive set — even by amounts that feel marginal — Amazon customers notice immediately. Review sentiment shifts. Conversion drops. Rank follows. You cannot out-market a pricing problem on Amazon. There is no endcap, no sales associate, no brand story that offsets a conversion signal.
The discipline required is to set pricing strategy before you need it — at channel entry — rather than reactively adjusting when contribution pressure arrives. Amazon should not be the channel where you try to recover margin from other accounts.
"Amazon doesn't care why your costs went up. It cares whether customers still click Buy."
6. Amazon vs. Retail: The Capital Allocation Question
Amazon and physical retail are not interchangeable channels. They are different businesses that require different resource allocations and produce different kinds of value.
Physical retail gives you visibility — brand presence, discovery through endcap and shelf placement, the associate recommendation, the tactile moment. Velocity is built more slowly, and the margin structure is different. But a strong physical retail presence builds a brand in a way that's hard to replicate digitally.
Amazon gives you velocity and direct feedback. The algorithm tells you, with almost no lag, whether your product is resonating. You can iterate faster, test pricing, and read demand signals in near-real time. But it builds almost nothing in terms of brand equity. Customers on Amazon often don't know or care who made the product — they know it had the best reviews and the best price.
The operators I've seen allocate across both well treat them as distinct P&Ls with distinct objectives. Amazon is the demand-capture engine — it serves customers who already know what they want. Physical retail is the demand-creation engine — it reaches customers who don't know yet.
When a brand shifts resources entirely to Amazon and exits physical retail, they often improve short-term contribution. Retail is expensive. But they also remove themselves from any space where brand equity is built and discovery happens organically. The channel mix question is ultimately a capital allocation question, not a marketing one.
7. The Operator Checklist: Are You Amazon-Ready?
Before treating Amazon as a primary growth channel — or before expanding your catalog there — these are the questions worth answering honestly.
Is your value proposition immediately obvious without explanation? If customers need to read three paragraphs to understand what the product does and why it's worth the price, Amazon will not do that work for you.
Can your pricing hold under sustained competitive pressure? Not for one quarter — for 24 months. If the answer is no, build that into your channel strategy before launch.
Is your supply chain built for the velocity targets you're actually projecting? Stockouts are expensive on Amazon in ways that don't show up immediately but compound over time.
Do you have a returns management strategy? A high return rate will eventually surface in review sentiment and eventually in algorithmic suppression. Know your expected return rate before you scale.
Are you watching category-level signals, or only your own metrics? Your rank and conversion can look fine while the category is shifting underneath you. The brands that get caught are the ones watching their own dashboard.
What is your Amazon strategy if the channel delivers 40% of revenue? The success case is also a risk. Concentration in a single channel — especially one where you have no direct buyer relationship — is a structural vulnerability.
The Operator Takeaway
Amazon is the most transparent channel in retail. The data is real, the feedback is immediate, and the algorithm doesn't lie. That transparency is an advantage if you know how to read it.
The mistake most brands make is treating Amazon like a retailer — building relationships, telling stories, expecting the early wins to compound automatically. The brands that operate Amazon well treat it like what it actually is: a continuous performance audit. Every day is a re-approval.
That is not a threat. It is information. Build your channel strategy around it.
#####
About Retail Readiness
Retail Readiness is a channel advisory practice built by operators, not consultants. Steven Bickers has 20+ years of CE retail channel strategy, including active account management at Walmart, Target, Best Buy, Costco, Amazon, and Sam's Club. Holly Sweezey is a former retail buyer at TJX, BoxLunch, and Crunchyroll — she sat on the other side of the table. Together they represent the full retail transaction: sell-side execution and buy-side decision-making.
"DTC proves demand. Retail proves discipline."