Food & Beverage brand

Consumable grocery is a category where the advertising maths is different from almost everything else on Amazon. The first purchase is rarely where the money is. What matters is whether that buyer comes back (through Subscribe & Save, through repeat search, or through habit), which means an acquisition cost that looks expensive against a single order can be entirely rational against a customer.

30-day reported performance

Ad sales	$66,578.15
Ad spend	$13,102.83
ACOS	19.68%
Purchases	1,854
Impressions	217,461
Clicks	3,115

Reading these numbers

The figures above cover a rolling 30-day window on the account’s Sponsored Ads.

19.68% ACOS on $66,578.15 in ad sales. Roughly one dollar in five of advertising-attributed revenue went back into advertising. For a repeat-purchase grocery product that is a comfortable position: profitable on first order, and considerably more so once repeat behaviour is counted.

3,115 clicks producing 1,854 purchases. That is a conversion rate near 60% on advertising traffic, which is unusually high and tells you something specific: the traffic being bought was extremely well qualified. Broad discovery was not carrying this account. Spend concentrated on terms where intent was already established.

217,461 impressions for 3,115 clicks. A click-through rate around 1.4%. In a category this visually competitive, that is a healthy signal on both main image and price positioning at the search results level.

What the structure did

Intent segmentation

Branded terms were separated from category terms entirely. In grocery, branded search volume grows as a direct function of repeat purchase. Leaving it mixed into category campaigns would have made the whole account’s ACOS look better than the acquisition work actually was, and hidden which terms were genuinely bringing in new customers.

Negation as routine

Consumable categories attract enormous volumes of loosely related search traffic: adjacent flavours, competitor formats, dietary terms that sound relevant and convert at zero. Weekly search-term work moved converting terms into exact match and negated the rest at the level that fitted the problem.

Placement discipline

Top-of-search placement was bid up where the data supported it and pulled back where product-page placement was doing the work more cheaply. In a category with high impression volume, small placement multiplier errors compound quickly.

1. 01Impressions217,461

   Times the ads were served

2. ↓ 98.57% did not click
   02Clicks3,115

   1.43% click-through rate

3. ↓ 40.5% of clicks did not convert
   03Purchases1,854

   59.5% conversion from click

Bar lengths use a logarithmic scale. On a linear one, clicks and purchases would be invisible against the impression count. The pale hairline inside each bar marks where that stage would actually end on a linear scale. The figures beside each stage are the reported values.

A 1.43% click-through rate in a visually competitive category is a healthy signal on main image and price positioning. The 59.5% conversion from click is the number that matters most here: it says the traffic being bought was already qualified, not broad discovery hoping to convert.

The constraint that shaped everything

Consumables have hard inventory limits. Expiry-dated stock cannot be over-supplied into FBA, which caps how aggressively any campaign is allowed to scale. Bid strategy was set against stock cover rather than against the ACOS target alone, pushing hard on products with runway, protecting rank rather than chasing volume on products without it.

This is the reason I do not run advertising separately from inventory. On this account they were the same decision, made weekly.

Related capability

Amazon PPC management and FBA inventory management.