221: Advanced Techniques With Amazon Sponsored Product Ads With Edward Ruffin

Podcast: Download (Duration: 50:28 — 92.4MB)

The single biggest mistake advanced Amazon sellers make with PPC is letting campaigns accumulate search term data and never acting on it. Edward Ruffin ran over 20 one-on-one account audits at my Sellers Summit, and the pattern he saw repeatedly was sellers with thousands of search terms sitting in their reports, plenty of them irrelevant, and no negative keyword cleanup happening at all.

Ed ran sponsored product ads at Seller Labs and manages high-volume client accounts daily. He is, by his own cheerful admission, a PPC nerd who thinks about ads from the moment he wakes up.

This episode skips the basics entirely. It covers his negative keyword process, why he bids at the top of the suggested range instead of the bottom, the 10-click rule for pulling keywords, how to tell an ad problem from a listing problem, and how to run dayparting tests using duplicate campaigns.

Some platform mechanics here have changed since this conversation. The Bid+ feature and the old headline search ad format have been replaced by Amazon’s placement bid modifiers and Sponsored Brands, and the underlying strategy in both cases still applies. Those sections are noted where relevant.

Key takeaways

• Run your first negative keyword cleanup after about one week of data, then re-check every few days on high-volume accounts.
• Use both negative exact and negative phrase match. Most sellers only use one.
• Bid toward the high end of Amazon’s suggested range or above it, then scale down. Starting low costs you two or three months of learning.
• Wait for 10 clicks on a keyword before making any judgment call on it.
• Showing up for wildly irrelevant search terms is a listing problem, not an ad problem. Amazon reads your listing to decide what you sell.
• Keep auto campaigns running permanently. They win placements manual campaigns cannot: related-product carousels, add-to-cart pages, and post-purchase pages.
• Pull the placement report to see which campaigns convert at top of page before raising placement bids, rather than turning it on everywhere.
• Start Q4 preparation in August so you have clean baseline data before traffic spikes.

How do you clean up Amazon PPC search terms with negative keywords?

Run your first cleanup about a week after launching a campaign, then continue checking on a regular cadence. Ed’s team lets new campaigns accumulate roughly a week of data, then pulls the search term report and negates everything obviously unrelated to the product.

The first pass is easier than it sounds. If you sell an apple slicer, terms like garlic press, keyboard, or Gatorade jump out immediately, and you do not need conversion data to justify negating them.

Amazon generates those odd terms deliberately. It is testing what traffic converts for your product, and your job is to shut down the tests that make no sense.

Use negative exact and negative phrase match together. Most sellers Ed audits use only one of the two, which leaves obvious waste running.

When should you negate a keyword that is relevant but not converting?

Negate a relevant term once it has accumulated clicks without conversions, since relevance alone does not make a term profitable. Ed’s example is “red apple slicer” driving steady traffic to an apple slicer listing that never converts.

That is a second, later pass. The first pass removes nonsense terms, and the ongoing passes remove terms that made sense on paper and failed in practice.

How do you tell an Amazon ad problem from a listing problem?

An ad problem means you chose unprofitable or overreaching keywords yourself. A listing problem means Amazon is serving you for terms you never targeted because it has misread what you sell.

The ad-problem version is a hunch without data behind it. You guess that someone searching “apple peeler” might also buy an apple slicer, and eventually the numbers show that people searching for a peeler want a peeler.

The listing-problem version looks different. You keep showing up for random unrelated searches, which means Amazon’s system parsed your title, bullets, and backend terms and concluded you sell something else.

Nobody at Amazon manually classifies your product. The algorithm reads your listing, so persistent irrelevant impressions are a signal to fix the listing rather than to keep adding negatives.

Should you bid high or low when starting an Amazon PPC campaign?

Bid at the high end of Amazon’s suggested range or slightly above it, then scale down as data arrives. Ed treats the suggested range as an estimate of typical bids rather than a rule, and deliberately overshoots it.

Speed is the reason. Bidding aggressively generates enough clicks in one or two weeks to make optimization decisions, where a cautious low bid means waiting a month or more for the same volume.

From that high starting point you have three moves. Lower the bid on keywords that get clicks and convert poorly, pull keywords that get clicks and never convert, and raise bids on the clear winners.

Starting low and creeping up takes two or three months to reach the same place. Your actual cost per click is often well below your bid anyway, since bids fall during periods when competitors have exhausted their budgets.

How often should you optimize Amazon PPC campaigns?

Check high-volume accounts every two or three days and steadier accounts about twice a week. Ed’s warning is against daily checking, which leads to overreacting to noise.

The failure mode is specific. A keyword gets 12 clicks and no conversions today, you pause it, and it would have produced 10 conversions tomorrow.

Always look at a two-week or one-month window rather than a single day or two. The longer range smooths out the variance and keeps you from making judgment calls that hurt you later.

What is the 10-click rule for Amazon PPC keywords?

Wait until a keyword has at least 10 clicks before deciding anything about it. Ed uses that threshold as his universal trigger for raising a bid, lowering it, or pausing the keyword entirely.

Ten clicks with zero conversions after a few weeks means the keyword is a candidate for pausing, since you have likely spent $20 to $30 proving it does not work. Ten clicks with six or seven conversions means raise the bid and stay competitive.

The middle case is the interesting one. Ten clicks with one or two conversions means the keyword is profitable but inefficient, so lower the bid to keep showing up at a lower cost rather than cutting it off.

Ed rarely pauses keywords outright unless there are no conversions at all. Scaling bids down is his default lever, since it protects ACOS without giving up visibility.

Impressions come last in his analysis. Being seen costs nothing, so impressions tell you about reach rather than about performance.

Should you run Amazon ads at a loss to build organic rank?

The answer depends on category and goal, and the endgame should always be profitability. Ed starts by calculating the client’s breakeven ACOS, then aims to at least break even while spend swells during the launch phase.

Some categories run at 100% ACOS as a norm. Supplements is his example, where sellers accept losses on the first sale because they are buying repeat customers and brand presence.

His own approach is to taper toward profit over time. Set an ACOS target, undercut it, and trim it down until the account sits in maintenance mode.

Why should you keep Amazon auto campaigns running permanently?

Auto campaigns win ad placements that manual campaigns cannot access, so pausing them after the research phase throws away inventory. Most sellers run autos briefly to harvest keywords and then archive them, which Ed considers a serious mistake.

The exclusive placements are valuable. Auto campaigns are what put your product in the related-products carousel on competitor listings, in the ads that appear when a shopper adds an item to cart, and on the post-purchase confirmation page.

Those post-purchase placements convert well. The shopper has already proven they will spend money on Amazon today.

Auto campaign clicks also cost significantly less than manual clicks, because placement depends on Amazon’s own traffic decisions. Ed suggests bidding $1 to $2 depending on category, watching the actual cost per click trend, then adjusting.

Running both campaign types together also serves brand recognition. Appearing in search, at top of page, on competitor listings, at add-to-cart, and at checkout builds familiarity that eventually converts on its own.

How do Sponsored Brands ads differ from sponsored product ads?

Sponsored Brands ads, called headline search ads at the time of this conversation, give you creative control that sponsored product ads do not: a custom headline, a chosen image, and multiple products in one unit. They require Brand Registry.

The image slot is the real advantage. Instead of a stock white-background product photo, you can run a lifestyle shot of someone using the product, a styled outdoor image, or your logo.

On mobile that image may be the only thing a shopper sees from your ad. Ed treats it as the single chance to pull someone in.

The headline gets roughly 50 characters and Amazon rejects claims you cannot substantiate. “World’s best” and any health claim will be denied, so keep it to a concrete product benefit or a seasonal angle like a Father’s Day gift framing.

Rejections are recoverable. Ed opens a case, gets the specific reason, fixes it, and resubmits successfully, and he recommends preparing five or six headlines so you can split test the ones that get approved.

How do you use Sponsored Brands ads to target competitor brands?

Target competitor brand names with a Sponsored Brands ad rather than a sponsored product ad, because the richer format gives the shopper a reason to consider you. A single product image competing against a brand someone specifically searched for rarely wins.

The Sponsored Brands version presents more. A lifestyle image, a headline explaining the benefit, and three of your products all appear at once, which reads as a genuine alternative rather than an interruption.

Expect lower conversion rates than your core sponsored product campaigns. You are reaching a wider and less qualified audience, and Ed’s stance is break even to modest profit on those campaigns.

Brand-name targeting is the opposite case. A Sponsored Brands ad targeting only your own brand name can run at a 2% or 3% ACOS, because everyone clicking already wanted you.

How do you use the Amazon placement report to decide bid modifiers?

Pull the sponsored products placement report to see which campaigns convert better at top of page, then raise placement bids only for those. The report splits performance into top of page and other placements over your chosen date range.

Ed’s earlier instinct was to distrust any feature that raises his spend. Backing the decision with placement data changed his mind entirely.

At the time of this episode the feature was called Bid+, which raised bids up to 50% for top-of-page and mid-page placement. Amazon has since replaced it with more granular placement bid modifiers, and the method is identical: confirm top-of-page conversion advantage in the report first, then increase the modifier for those campaigns.

Seasonal and single-day sales are strong use cases. If you make most of your revenue in one quarter or one promotional day, buying top-of-page visibility during that window is worth the premium.

How early should you start preparing Amazon ads for Q4?

Start in August so you have clean baseline data before holiday traffic distorts everything. Ed wants to know normal traffic patterns, confirm which search terms his auto campaigns surface, and verify his manual campaigns target converting keywords before Q4 arrives.

Building sales velocity ahead of the season also helps organic ranking. Launching a product during the holidays is much harder, since you are still trying to index for keywords while competing against peak traffic.

Listing optimization belongs in this window too. Fixing a listing that has Amazon serving you for the wrong category is far cheaper before the traffic surge than during it.

How do you adjust Amazon ad budgets and bids for the holidays?

Double or triple daily budgets and roughly double bids going into peak periods, so campaigns do not exhaust their budget mid-day. An ad that runs out of budget stops being seen at exactly the moment traffic peaks.

Higher spend is the point, and clients need to understand that going in. Conversion rates rise because shoppers are in a buying mood, so the increased spend converts at a better rate than usual.

This is also when the placement report earns its keep. Campaigns that already convert well at top of page are the ones to push hardest during the holidays.

What is Amazon PPC dayparting and how do you test it?

Dayparting means running campaigns only during the hours of day that convert best for your product, rather than letting them burn budget around the clock. Testing it requires duplicate campaigns split across time slots.

The test method is mechanical. Create the same campaign 12 times covering two-hour segments, or 24 times covering one-hour segments, and after about two weeks the data shows exactly when your product converts.

Ed’s observed pattern across accounts is that budgets reset at midnight Pacific and most sellers exhaust theirs by 9am Pacific. That means the afternoon and evening hours, roughly 3pm through 7pm Pacific when people are home from work, often see the strongest conversions.

Cheap clicks are the second benefit. Once competitors run dry, bids drop sharply, and Ed has picked up clicks at 15 to 20 cents that produced sales at a 1% ACOS.

That is another argument for always-on auto campaigns. They are positioned to absorb the traffic nobody else is bidding on late in the day.

Product category shapes the pattern. Coffee might sell in the morning, or at night when someone realizes they are out for tomorrow, and only the test tells you which.

Frequently asked questions

How many clicks before you pause an Amazon PPC keyword?

Wait for at least 10 clicks before making any decision on a keyword. Ten clicks with no conversions over a few weeks justifies pausing it.

Should you use negative exact or negative phrase keywords?

Use both. Negative phrase catches variations around a term, negative exact blocks a single specific query, and most sellers apply only one and leave waste running.

How often should you check Amazon PPC campaigns?

Every two or three days on high-volume accounts, and about twice a week otherwise. Checking daily leads to overreacting to normal variance.

Should you start with high or low bids on Amazon?

Start at the high end of the suggested range and scale down. It generates usable data in one or two weeks instead of two or three months.

Why does my Amazon listing show up for unrelated search terms?

Amazon reads your title, bullets, and backend terms to decide what your product is, so persistent irrelevant impressions mean the listing is miscommunicating. Fix the listing rather than only adding negatives.

Should you keep auto campaigns running after finding your keywords?

Yes. Auto campaigns win placements manual campaigns cannot access, including related-product carousels on competitor listings, add-to-cart pages, and post-purchase pages, and their clicks cost less.

What ACOS should you target on Amazon?

Start by calculating your breakeven ACOS, then set a target below it and trim toward it over time. Some categories such as supplements tolerate 100% ACOS because sellers are buying repeat customers.

When should you start preparing Amazon ads for the holidays?

Start in August. You want baseline traffic data, clean search term targeting, and an optimized listing before Q4 volume arrives.