Amazon Dynamic Bidding and Placement Multipliers in India

The bid stack is a product, not a sum

Most Amazon India accounts we audit run one bidding strategy across every campaign because someone chose it at setup and never revisited it. That is expensive. On Sponsored Products, the price you can pay for a click is built from three layers that multiply together: your keyword or target bid, your placement multiplier, and the dynamic adjustment Amazon applies in the auction. A Rs 12 bid is almost never Rs 12.

The order matters. Amazon applies the placement multiplier to your base bid first, then applies the dynamic adjustment on top of that result. Set a base bid of Rs 12, a Top of Search multiplier of 100 percent, and the strategy to dynamic bids up and down, and your ceiling for a top of search click becomes Rs 12 x 2 x 2, which is Rs 48. A seller who sets a 300 percent multiplier to be safe on a Rs 20 bid has authorised Rs 160 clicks in a category where average order value is Rs 900. The auction rarely charges the ceiling. But the ceiling defines the worst case, and in festive weeks the worst case shows up far more often than it does in a quiet Tuesday in June.

Three strategies, three different jobs

Down only lowers your bid in real time when Amazon judges a click unlikely to convert, and never raises it. It is the default for new campaigns and the right setting for defensive brand campaigns and for auto campaigns you run purely as keyword harvesters. It cannot overspend you, but it can quietly starve you of impressions on high intent queries.

Up and down can raise your bid by up to 100 percent for Top of Search and up to 50 percent for other placements when conversion looks likely. This is the only strategy that lets you win a slot you would otherwise lose, and it is also the fastest way to double a CPC without noticing. It earns its place only where you already have a stable conversion rate on at least 100 to 150 clicks.

Fixed bids apply no adjustment in either direction. Treat this as an instrument rather than a setting. When a campaign with down only shows falling impressions and rising ACOS at the same time, switching to fixed for ten days tells you whether Amazon was suppressing your delivery or whether the keyword genuinely stopped converting. Fixed is also sensible for a launch window where you want predictable delivery rather than optimised delivery.

Read the placement report before touching a multiplier

The placement view in Campaign Manager splits performance into Top of Search, Rest of Search and Product Pages. Pull 30 days per campaign and put spend, clicks, orders, conversion rate and CPC side by side. The pattern we see repeatedly in Indian categories is a Top of Search conversion rate that runs 1.5 to 2.5 times Rest of Search, against a CPC that runs 1.8 to 3 times higher. Those two ratios are the whole decision.

The test is simple. Bidding up on Top of Search only pays when the conversion rate ratio is larger than the CPC ratio. Take a real shape from a home category account: Top of Search converts at 4.1 percent with a Rs 21 CPC, Rest of Search at 2.2 percent with a Rs 9 CPC. The conversion ratio is 1.86. The CPC ratio is 2.33. On that keyword, a Top of Search multiplier destroys margin even though the placement looks like the best performer in the dashboard.

Run this per campaign, not per account. Placement economics differ wildly between a branded defence campaign, where Top of Search converts three times better because the shopper typed your name, and a generic category campaign, where Top of Search buys you a browsing audience that was never close to a purchase.

A decision rule that holds up across categories

• Branded defence campaigns: fixed bids, Top of Search multiplier between 25 and 50 percent, and a hard bid cap set at the CPC that keeps ACOS under half your contribution margin.
• Generic non-brand campaigns: down only until the campaign has 200 clicks, then apply the ratio test before any multiplier.
• Auto and broad harvest campaigns: down only, small fixed budget, judged on the search terms they surface rather than on ACOS.
• ASIN and product targeting campaigns: use the Product Pages multiplier, never the Top of Search multiplier, because these ads mostly serve on detail pages.
• New launch campaigns: fixed bids for the first two weeks so that delivery is stable enough to read.

How to change bids without destroying the read

Two habits separate accounts that improve from accounts that just churn. First, change one variable per campaign per cycle. If you raise a base bid and add a placement multiplier in the same week, you will never know which one moved the number. Second, respect the attribution window. Amazon credits conversions to clicks for fourteen days, so a bid change contaminates the next fortnight of reporting. Use four week rolling comparisons and ignore the first seven days after any change.

Budget caps quietly ruin all of this. A campaign that exhausts its daily budget by 2 pm is not being tested, it is being rationed, and its placement mix reflects whatever the morning traffic looked like. Before comparing placements, confirm the campaign ran the full day for the period you are analysing. Fix pacing first, then read performance.

What we would do in the first two weeks

Week one is diagnosis only. Export 60 days of placement data for every campaign, calculate the two ratios per campaign, and mark every campaign where an existing multiplier fails the test. Also list every campaign that hit its budget cap on more than five days, because those reads are unusable. Change nothing yet except the pacing problems.

Week two, act on the top ten campaigns by spend. Remove multipliers that fail the ratio test, move mature branded campaigns to fixed bids with a modest Top of Search multiplier, and switch any campaign with more than 150 clicks and a Top of Search conversion rate well above the campaign average to up and down. Then leave it alone for a fortnight. The discipline of not touching bids for fourteen days is worth more than any single bid change you might make in that window.