USE AMAZON SPONSORED

PRODUCTS ADVERTISING TO
HONE YOUR KEYWORDS
Bernie Thompson
For online marketplaces today, it’s only viable to skip in-market advertising if your
product is truly unique and you have some other way of making people look for it
specifically in organic search. Or, if you already have brand awareness on the level of
Apple iPhones.

Since neither is true for most of us, expect to advertise for most, if not all, of your
products. Two key ways paid promotion can help you:

1. Gain search insight. Paid promotion is the best way to find out what potential
customers search for, and how. Amazon provides advertisers with really good data
about whether those keywords were right for that product — if people did in fact
search, click, and buy. Do they type “luggage with locks” or “TSA-proof suitcase?”

2. Increase relevance. Remember, you need to train Amazon’s artificial intelligence

that your product is a high converter for given keywords. Paid ads, subsequent
shopper clicks, and then sales create evidence that your product should be a trust-
ed, relevant result for those keywords. Ad clicks that don’t generate sales will hurt
your conversion rate and represent lost advertising pennies. So select the keywords
you bid on for advertising thoughtfully.

There’s a third benefit, too: It’s impossible to understand, let alone influence, the
interactions and dependencies between advertising keywords and organic search
unless you take them together.

Once you’ve got some likely keywords — both broad, such as “pet supplies,” and
narrow, such as “heated dog bed,” plan your keyword campaigns with the following
steps.
 GET READY BY
IDENTIFYING METRICS
Smart advertising decisions require a few data ingredients:

Know the product’s profit margin. Keep direct and category fees in mind to
ensure that a campaign doesn’t lose money wildly. You may be willing to forgo profit
for a while to meet other goals, but don’t forget why you’re in business.

Know the organic ranking performance to start. A huge benefit of Sponsored

Products advertising is its positive effect on the product’s category ranking and rank-
ing in organic searches on the same keywords. Comparing those rankings before
and during a campaign can help demonstrate whether that campaign is a success.
(Tracking tools that can report ranking by keyword can help.)

By the same token, don’t bid a lot for keywords where you’re already ranking well —
in those cases, the flywheel is already working for you, so it’s not the best use of your
advertising dollar.

Identify campaign goals and a target ACOS. Are you trying to make money? Raise a
new product’s visibility? Own a particular keyword to improve organic search place-
ment? The goal will influence the advertising cost of sales (ACOS) you’re willing to
incur, and for how long. Note that ACOS doesn’t include direct and category fees,
which is why you need to know your margin to determine a wise ACOS.

THE TOP CAMPAIGN

METRIC: ACOS
To use paid ads effectively, pay at-
tention to each product’s advertising
cost of sales (ACOS). This term,
which is more common in the
Amazon world than elsewhere, is the
cost of a given sale attributable to
the advertising. You can obtain this
metric from the Campaign Manager
interface. It’s also provided in the
performance reporting of the
Amazon advertising application pro-
gramming interface (API,) which can
save time for sellers who manage
large or complex campaigns and
who understand the use of APIs. 1
In traditional markets, you’d rarely allow advertising costs to eat up your entire profit
margin, but in online markets, the calculation can be more complex. For new prod-
ucts or those you believe in but that aren’t getting traction, plan to incur a high ACOS
to get the flywheel spinning. It’s well worth the investment. Depending on your mar-
gins, consider spending enough for an ACOS of from 20 to 40 percent.

Once the product is established and doing well, you’ll want to keep ACOS relatively
low. So scale back to a percentage above zero but less than you’d spend at launch.
The value of PPC campaigns can vary by category, so experiment with the percent-
age you need to keep sales momentum from dropping off.

Determine your goals at the keyword level; a given ACOS percentage may represent
a huge success for a popular keyword, for which competition is tough, but the same
figure may be a poor performance for a long-tail keyword you should be able to own.
Incorporate realistic insight about competitive products that rank organically higher
than yours. What can you learn from those sellers? Does their product actually have
more features, a better price, or characteristics that relate better to a given keyword
than yours? If so, recognize that you may not ever win that top bid or top ranking.
That doesn’t mean you can’t succeed; just plan your goals accordingly.
 USE AUTOMATED
KEYWORD CAMPAIGNS
APPROPRIATELY
Sellers can choose from two types of Sponsored Ad campaigns: automatic and
manual. In an automatic campaign, you empower Amazon to cast a wide keyword
net for a maximum advertising spend. Automatic campaigns have very poor conver-
sion rates, but starting here does have two big benefits:

When you’re launching a product, it can help identify the category’s top successful
keywords, as well as no-brainers you may have missed.

The targeting report you’ll receive from automatic campaigns is the only way to see
real customer searches. This report is very helpful for later setting up effective
exact match and phrase-match bids for your manual campaigns.

AMAZON KEYWORD REPORTS

As of this writing, Amazon provides two reports relat-
ed to keywords. Their API names are the Sponsored
Products “Mega Report” and the “Automatic Target-
ing Report,” a.k.a. the search term report. (The
names are slightly different in the online seller control
panel.) The most important data to pull from these re-
ports are what keywords shoppers are actually using
to search, conversion rates by keyword, and trends.

If you’re new to PPC, use an automatic campaign like fishing bait to test your instincts
and catch bigger, better keyword combinations. Review the reports to see what
shoppers were searching for when matched to your bids. Draw more or better key-
words from that data. (See the case study at the end of this chapter for examples.)

TIP: CHECK YOUR CATEGORIES

When you’re selecting ad keywords, plug them into an Amazon search. The top five sub-
categories that appear under “Show results for” are the most relevant.

If your product is not listed within one of those five, your ad is unlikely to be displayed, re-
gardless of your bid.
 MOVE TO MANUAL
CAMPAIGNS & ESTABLISH
BIDS BASED ON
TARGET METRICS
Once you’ve identified some keywords that work, create manual campaigns with
those keywords and variations of them, perhaps starting with bids on broad matches
and phrase matches.

Keyword variations are very important for phrase-match and exact-match types, and
you can expect to add variations throughout the life of the product. Initially, the auto-
mated campaign will provide the bulk of variations and long-tail keywords, but even-
tually the broad-match and phrase-match words in your manual campaign will
become a better resource for variations. (That’s because by then, you’ll have more
search data for your specific product than Amazon, which can only pick keywords for
your automated campaign from those in your listing and category — not everything
users actually use to search.) You may wish to continue the automated campaign for
a while, even as you begin a manual campaign, to continue harvesting keywords. At
some point, the automated campaign will stop finding additional variations, and then
you can retire the automated campaign for good.

Some sellers include top competitors’ brand names in their advertising keywords. As
noted in the section on identifying product keywords, we don’t recommend this and
Amazon considers it a policy violation. A shopper who enters the brand name has
pretty good idea of what they want, so you’re fighting an uphill battle to get them to
click, let alone buy. For advertising, the ACOS is probably not worth it, and neither is
an account suspension.

For all of your keywords, base your bids on data, not the unreliable information
Amazon provides to suggest what a winning bid might be.

If your product already has some or-

ganic sales, take the Detailed Page
Sales and Traffic Report and look up
your unit session percentage, which
is essentially the conversion rate
once a shopper has viewed your
product page. If your keywords are
reasonably valid, this organic conver-
sion rate should be fairly close to
your ad’s conversion rate. Use that
rate and do the math to identify a
maximum bid that will achieve your
target ACOS (and acceptable profit
margin). Of course, the tools avail-
able can speed that calculation.
Unit session percentage indicates the product page’s conversion rate.
 If you’re building visibility for a new product without many organic sales, you won’t
have enough unit session data to be as certain about your keywords or the right
bids. Experiment, and expect to spend more time on the campaign than for an es-
tablished product until you have better data.

Open your bids on the high end of your acceptable ACOS. As long as you outbid the
next highest bidder, you may not pay your full bid per click anyhow. But note that you
can bid well above Amazon’s suggested high bid and still not get ad placement be-
cause conversion rates and maximum spend are also factors.

You can also bid the minimum of only two cents but still have your ad shown, if
Amazon is impressed by your conversion rate and/or if all your competitors for that
keyword have met their maximum spends, possibly by paying too much for other
keywords that don’t have a high probability of converting.

In the middle of that bidding spectrum are lots of sellers with products that convert
moderately well and who will pay quite a bit for an ad. Position yourself on the
low-cost end of that range, paying relatively little but with such a strong likelihood of
a sale that Amazon will show your ad.

How do you locate that sweet spot? Experiment.

 RUN SMALL...
CONTROLLED TESTS
Bid on a limited number of keywords for short periods. Keep close tabs on impres-
sion and conversion rates and ACOS to determine whether each $1 spent yields
$100 in sales — or zero.

Top keywords in a category may get three or four times the impressions of a lesser
keyword. What matters most, though, is sales. Identify which keywords get you sales
and at what ACOS. Depending on your goals, you might be temporarily okay with a
high ACOS or even a loss to get the flywheel going. On a great new product, this
strategy could give it dominance in organic searches for that keyword so you could
make up for lost profit in the future. But you can’t advertise away your margin indefi-
nitely.

TIP: SET UP NEGATIVE KEYWORDS, TOO

Don’t forget to include negative keywords in your campaigns. Protect your conversion
rate (and make the entire campaign more effective) by ensuring that people who are
clearly not after your product won’t be shown your ad by mistake.
 GET TOOLS THAT
ENSURE YOU CAN USE
YOUR RESULTS DATA
Amazon provides gallons more data about your keyword performance than Google
does. You’ll receive weekly reports with perhaps 100,000 rows of data about what
people are using to search and whether they’re buying.

You can drown in that much data without good tools to help you act on it. The ideal
tool summarizes the data to actionable steps, ranks keyword quality, and alerts you
to those you should do something about. By distilling the myriad choices into a few
big levers to pull, you’ll dodge the time-suck associated with keyword ad tweaking.

In fact, this might be the most important tool you can use, because without a good
tool to support proactive campaign management, you can spend money on ads that
don’t help as much as they should. Consider this “before and after” data from an ad
campaign for a single product. The top chart captures data from a combination of au-
tomatic and manual campaigns that were set and ignored. It shows that, while the
advertising increased sales, it had little effect on overall session percentage (conver-
sion rate) or, more importantly, organic sales. Total sales and ad sales are nearly iden-
tical. That’s an indicator that the campaign wasn’t affecting search relevance as
much as it should, which in turn meant that the keywords in the campaign weren’t
ideal.

On the other hand, when the

campaign was adjusted and
managed based on the sugges-
tions in this chapter, including
reducing bids on ineffective key-
words and increase phrase and
exact-match bids, organic sales
grew significantly, as shown
below. This was because the ad
click-throughs and sales in-
creased the product’s relevance
for those effective keywords, so
it also began ranking more
highly in organic searches for
those keywords, and that higher
search placement prompted
more organic sales, too. That’s
the flywheel effect, with paid
clicks increasing relevance and
search placement for organic
searches.
 INCREASINGLY FOCUS
ON LONG-TAIL KEYWORD
COMBINATIONS
Once you’ve started to identify successful keywords, focus your advertising spend.
This doesn’t necessarily mean higher bids on bigger keywords.

For any product, some keywords are more likely to lead to a sale than others — and
some are better for Sponsored Products advertising. Counter-intuitively, these two
groups are not identical, because the keywords most likely to result in a sale may also
be the most expensive, so advertising on a weaker keyword can be a better invest-
ment.

True, a lot of people search using broad, high-level keywords like “bicycle” or
“T-shirt.” The problem is that many sellers are competing for those keywords, so
they’re costly, and your chances of winning the bidding is small. When you do, your
ad may be seen by more shoppers, but a rather small subset might actually want
your specific product. On the other hand, narrower, “long tail” keywords are much
more available while also potentially more targeted to your ideal buyer. (Plus, as noted
earlier, amazing numbers of people perform unique searches with those long tail
terms. As a result, being the winning bidder on more small and inexpensive keywords
is potentially far more effective than winning a very few big keywords.)

So instead of a broad keyword like “laptop,” bid for “laptop touchscreen 8 GB RAM.”
The ad will cost less and the shopper will be more likely to buy because she already
knows exactly what she wants. That means the feedback loop of “ad shown > shop-
per click > sale” becomes self-reinforcing, boosting your product’s relevance in or-
ganic searches on those terms, too.

Continually expand your keyword selection and variations with new keywords that are
narrow but still relevant. Fewer people will use them to search, but they’re both less
costly and more likely to lead to conversions. Pause the campaign as needed to
make adjustments, and keep records to guide future campaigns. There are so many
variables, including changes in Amazon’s algorithms, that the “right” keywords will
continually shift, so don’t expect to stick with even good keywords forever. Be pre-
pared to manage them continuously.

HOW MANY KEYWORDS PER CAMPAIGN?

Once you begin using exact-match bidding, the need to include multiple combina-
tions of even a few words — like “best women’s raincoat” and “raincoat women’s
best” makes the keywords add up fast, sometimes to several hundred or more. A
campaign of 30 keywords would be fairly small. Regardless, adjust them as the cam-
paign runs to make it as effective as possible. Sellers should expect to add variations
throughout the lifespan of the campaign (and the product, if, like most, it needs an
occasional boost from advertising).
INCREASINGLY ADD
EXACT MATCHES TO
YOUR CAMPAIGN
It’s useful to begin campaigns (or add new, experimental keywords) with phrase
matching and use the resulting reports to identify the actual search combinations
shoppers use most frequently. But for any given keyword, exact-match bids will beat
phrase and broad-match bids every time. So as you identify your most effective key-
words, increasingly add exact-match bids on the most effective combinations.

Exact-match bids are the most restrictive, but they’re also likely to attract your most
targeted customers, for the same reason as narrower keyword combinations. That’s
good for conversion rates, and in turn more impressions and better organic search
placement. Exact-match keywords may also be less expensive because there’s less
competition for them.

THE THREE BID TYPES

Suppose you’re bidding on a keyword combination of “bottled alkaline water.” You’ve
got three choices for bidding:
Exact-match bids for the entire keyword combo. The shopper must type “bottled
alkaline water,” fully and in that order, no more and no less, for your bid key-
word to match.

Phrase-match bids. In this case, the shopper could type “six-pack bottled alkaline
water” and your product would still be suggested. Plurals are usually overlooked,
along with additional words on the front or back, so “six-pack bottled alkaline
waters” would match your keyword, too. But the search terms still have to be in
the same order as your combo, and other words can’t be inserted be-
tween them. “Alkaline water bottled” would not match, and neither would “bot-
tled six-pack alkaline water.”

Broad-match bids, which in addition to singulars or plurals, allows words to be

reversed, omitted, or added, and admits both some spelling variety and syn-
onyms. Using the example above, the shopper could type “spring water, alkaline,”
or any of the non-matching phrases above, and your product might still be sug-
gested.2
 ADJUST YOUR AD SPEND TO
MEET YOUR ACOS TARGETS
Once the campaign has run for at least two weeks, lower your bids on any keywords
that get impressions but don’t convert very well (and thus incur a higher ACOS than
expected). Then any subsequent impressions and sales will better meet your ACOS
targets. If that doesn’t work, remove those weaker keywords.
Start with automated campaigns to gain insight and keywords, then move to
manual campaigns you can tweak as more data comes in.

If you’re not getting many impressions for a keyword, assess whether it’s simply not
a great match for your product or your bid is too low. You can find out for sure by bid-
ding significantly higher. If more impressions result, there’s your answer—though
you’ll still need to decide if it’s worth the ACOS.

If you want to spend less in general, remove broader keywords, which may convert
well but probably at higher costs. When you want to spend more, increase the bids
for your strongest keywords — especially those that result in a low ACOS. Or add
new (and preferably narrow) keywords.
BIDDING MORE — OR MORE WIDELY?
It can be hard to decide when to increase bids and when to add keywords. Amazon
will tell you what a winning bid would be, but this amount is probably an average (and
not the only factor in the system’s ad placement decisions), so sellers often have to
bid well above that amount to actually win good placement.

If you bid a little below that theoretical winning bid, particularly if your product has
been converting well, you’ll still receive impressions — sometimes plenty of them —
just not as many as the winner. That’s one case in which you might want to increase
your bid slightly to see how ACOS changes. If you’re still struggling to win impres-
sions, however, you may need to bid much higher to get them. Only experimentation
can prove whether the ACOS is worth it.

In addition, in 2016 Amazon introduced “Bid Plus,” in which a seller gives permission
to spend 1.5 times their bid to get top positioning on the results page. It’s impossible
to win that top spot without it. Using Bid Plus and tracking the results can be another
way to determine if an increased base bid can put you among the frontrunners to
make the increase worthwhile. If your bid is a long way from the winner, it’s generally
more effective to add keywords instead. See the troubleshooting section of this guide
for more information.
USE YOUR AD DATA TO TWEAK
THE PRODUCT LISTING
Remove weak keywords from your most important keyword sources, the title and
product bullets. Make sure those sources include your strongest keywords promi-
nently. Add new, long-tail keywords to your hidden keyword lists to leverage your ad
campaigns and further improve your organic search results.
USE THE DATA TO DIRECT
OTHER MARKETING
Take your successful keywords to other marketing channels.
Sponsored Products isn’t your only Amazon option. There are others, some of which
compete for the same space on search pages. For instance, the options of Amazon
Marketing Services — headline search ads, product display ads, and a brand page
— can be useful, but as of this writing, those campaigns can only be manipulated by
hand and offer more limited results data. So gain confidence about your keywords,
the conversion power of your product page, and how to design successful ads
before taking that step.

Another option is Amazon Business. This B2B version of the marketplace is a

low-overhead solution for buyers whose volumes warrant a discount or other special
arrangements but aren’t large enough to buy directly from you. (See the inventory
management chapter of this guide for more details.) The keywords that work well on
Amazon Business may be different than for the regular marketplace, but experience
and data from one can inform the other.

All of these special programs can help drive traffic to your product and bolster con-
versions, which then improves your discoverability everywhere. As you get your prod-
uct pages, keywords, reviews, and Sponsored Products ads working well, check out
other promotional options to boost your flywheel momentum.

Similarly, when you’ve identified low ACOS keywords, advertise with the same key-
words on Google, Facebook, or any other online channel.

1
www.advertising.amazon.com/API
2
Sources for these distinctions: www.ams.amazon.com/faq and http://www.salesbacker.com/blog/65/Broad__Phrase__or_Exact_-
Match__Which_Amazon_PPC _Search_Type_Should_I_Use_

ABOUT EFFICIENT ERA

Efficient Era is an all-in-one suite of AI-powered tools for Amazon sellers, designed to help you increase
sales and save money. Our Ad Manager automatically discovers new keywords and recommends bid ad-
justments, saving you thousands of dollars you'd otherwise spend on expensive PPC consultants or tools.
Our business dashboard helps you measure profitability and eliminate unprofitable products. But that's not
all. Affordably priced at just 2 cents per order, this essential collection of over 20 different tools, including
feedback emails, customer insights, and BuyBox alerts, makes signup a no-brainer!

Compiled by Efficient Era.