Ask store owners how they found their keywords and many describe a single step. They open Keyword Planner, paste in their own website address, and export whatever comes back. It feels thorough: thousands of rows, real volumes, Google's own data.
We tried exactly that for three online fly fishing shops that sell Simms waders. Keyword Planner returned thousands of ideas for each one. "Simms waders" was missing from all three. Americans type it about 12,100 times a month, the same as "fly rod."
This chapter covers keyword research for an online store the way we do it: see who's already advertising, seed Keyword Planner with more than your own site, clean the list before a dollar goes out, and launch the searches you're sure of on exact match. The examples come from the coffee, e-bike and fly fishing stores in this guide. Your products will differ. The process doesn't.
Keyword Planner Only Knows What You Feed It. Your Own Website Is the Narrowest Seed There Is.
Start before Keyword Planner. Google's Ad Preview and Diagnosis tool shows you a live results page for any search, as Google would show it, without affecting your impression data. Type in your main product searches and note who advertises: other stores, the brands you sell, marketplaces. Resist clicking their ads. The preview already showed you what you need, and the click lands on a rival's bill.
Then open Keyword Planner's "Discover new keywords." Google lets you start from words, from a website, or both. Many store owners we talk to only enter their own site, and that's where the trouble starts. In our experience, Keyword Planner leans hard on what your pages say and suggests searches close to that wording. If your product pages lead with your own product names and lifestyle copy, the suggestions follow your copy instead of your customers.
So we seed it three ways:
- Your own site, to see what Google thinks you sell.
- A rival store's site, one you found in the preview, to see the searches you're missing.
- The words your buyers use: product types, materials and, for a reseller, the brand names on your shelf.
A word on the bid columns. Each idea comes with a low and high top-of-page estimate. We treat both as ballpark figures that can miss by roughly 30% either way; they forecast an auction, they don't price your clicks.
"Simms Waders" Draws as Many Searches as "Fly Rod." Fly Shop Buyers Search the Brands on the Shelf.
Our three fly shops are resellers. They carry rods, reels, waders and flies from brands like Simms, Sage, Orvis, Redington and Echo, and each one sells the Simms G3 Guide waders. A real share of their customers, roughly 10% to 24% of the buying searches, type the brand, often with the model.
When we measured the site-seeded Keyword Planner pulls for the three shops, searches naming a manufacturer brand made up 24.0%, 21.4% and 9.5% of the buying searches. The shops' own names were around 1% or less. And those pulls still missed some of the biggest brand searches. A direct check found "simms waders" at 12,100 a month; even a single rod model, "sage r8 core," showed 590.
The fix is simple and easy to skip. Type the brands you carry into Keyword Planner as seed words, one brand at a time, with the product type: "simms waders," "sage fly rod," "orvis reel." Our brand check turned up a ready-made ad group in one pass: "simms waders" (12,100 a month), "simms g3 waders" (1,300) and "simms g3 guide waders" (480), all wanting the same Simms wader page. "Patagonia waders" (6,600) and "orvis waders" (3,600) each earn a group of their own, pointed at that brand's waders. Then sort what comes back by brand and product. Each brand-and-product pair is a natural ad group, and each one should land on a page that shows that brand's version of that product.
Using a brand's name is allowed for a reseller. Google's trademark policy lets ads use another company's trademark when identifying a product for sale, as long as the landing page sells that product and makes clear you're a reseller. A brand owner can still complain about ads that use its name in a confusing or misleading way, and some dealer agreements add their own terms, so check yours.
A direct-to-consumer brand that makes its own products has a different problem: few people search its name yet. For a store like that, the seed words are product types and materials, which Chapter 4 covered, not brands.
One Coffee Roaster's Website Pulled In Searches for Coffee Shops, Not Coffee Beans. Sort the List Before You Spend.
Seeding from a site can also bring back too much. When we seeded Keyword Planner with Onyx Coffee Lab's website, a roaster whose site also promotes its cafe, most of the raw volume was searches for coffee shops, cafes and "near me." Even after we removed the two giant "coffee shop" rows, the list was still 18.7% place searches, 17.2% equipment like grinders and brewers, and 7.0% cold-brew recipes and how-tos.
None of that is someone looking to buy a bag of beans online. The other two coffee stores' pulls were much cleaner, with place searches under 2%, because their sites read like stores, not cafes. Keyword Planner reflects what your website talks about, so a store whose site also covers a cafe, classes or a blog will get those searches back.
Clean the list before it becomes an ad group. Sort every idea into buckets (product searches you sell, product searches you don't, places, equipment, jobs, how-to) and keep only the first.
It goes faster than it sounds. Sort the export by volume, read the top few hundred rows, and the noise falls into a handful of patterns: a city name, "near me," a machine brand, "recipe," "jobs." Filter on those words and most of the junk leaves in a few passes. What's left is a list a shopper with a credit card would actually type.
"Cheap Electric Bike" Is a Negative for One Store and the Whole Market for Another.
Some words almost always mean "not a buyer." In our e-bike pulls, "free ebike" draws about 4,400 searches a month. Nobody typing that is about to buy a bike today. "Used e bike," about 12,100 a month, belongs on the negative list for any store that only sells new bikes. "Coffee roaster jobs," 880 a month, is someone looking for work.
Other words depend on what you sell. "Cheap electric bike" draws about 110,000 searches a month. For a store whose bikes start around $3,000, those are window shoppers who'll leave at the price. For a store whose ads say its bikes start at $799, as one of our e-bike stores' ads did, that search may be the heart of the market.
So don't copy a negative list from another store, or from a blog post. Decide each word against your own prices and products.
One Google detail matters here: negative keywords don't match close variants. If you block "cheap," Google won't automatically block "cheapest" or "cheaper." Add the versions you see.
A Rival Store's Name Goes on the Negative List. The Brand on Your Shelf Doesn't.
When you seed Keyword Planner with a rival's site, its store name comes back in the ideas, sometimes in large numbers. Later, when broad match runs, rival names show up again in your search terms. We put rival store names on the negative list. Someone searching for a specific store usually wants that store, and paying to appear for its name is a separate decision, one we take up with brand campaigns in Chapter 12.
Manufacturer brands are different. "Simms waders" isn't a search for a rival; it's a search for a product a fly shop sells. Treat it like any other product search. Orvis is the tricky one: it's a brand fly shops carry and also a store with its own website. Searches for an Orvis product are fair game for a shop that stocks it; searches for the Orvis store are not.
The negative list is never finished. Start it on launch day with the obvious ones from your research: free, used, jobs, repair, rival store names. Then let the search terms report add the rest. It grows the first week you launch and keeps growing, fastest once broad match runs. That's when search terms need a daily look.
Exact Match Goes Live First. Broad Match Waits Until the Account Makes About 30 Sales a Month.
With the list clean, build the groups from Chapter 4: a handful of near-identical searches, an ad written for them, and the category or brand page that answers them. Then launch on exact match with bids set by hand. We call that campaign Isolation: exact match only, so every dollar goes to searches you chose.
Broad match comes later. Around 30 purchases a month is our trigger, which a store often reaches a couple of months in, and busy ones sooner. At that point we open Exploration, our label for a broad-match campaign whose bidding chases a target. It goes looking for searches no keyword list would include, and its search terms are where your negative list grows fastest.
Next: Standard Shopping Still Takes Manual Bids, and That Makes It Your First Shopping Campaign.
Keywords cover Search. Chapter 6 adds Shopping, the product ads with photos and prices, and makes the case for Standard Shopping with bids set by hand as the store's first Shopping campaign.
Earlier chapters cover why category pages win the auction and what a click can cost before it loses money. Every chapter is on the guide's home page.
When we build keyword lists for stores through Google Ads management, the brand names on the shelf go into Keyword Planner before the store's own website does. Doing it yourself? Seed Keyword Planner with your top three brands or product types and compare the list to what your website alone returned.
