Building a Keyword Research Strategy From Scratch
How to build a keyword research process in 2000 using WordTracker's query data and GoTo.com's bid listings, before dedicated SEO tools existed.
Keyword research in 2000 doesn’t look like a spreadsheet exercise yet, because the tools that would later define the discipline are brand new. WordTracker, launched in 1998, is one of the first services built specifically to show what people are typing into search boxes. GoTo.com’s public bid listings reveal something just as useful: what advertisers are willing to pay for a term, which is a rough but honest signal of commercial intent. Here’s how to put both to work.
Step 1: Start with your own product language, not guesses
Before touching any tool, write down the words your actual customers use when they describe what you sell, not the industry jargon your business uses internally. Ask a few customers how they’d search for you if they didn’t already know your business existed. This list is your seed set.
Step 2: Check real query volume with WordTracker
WordTracker pulls its data from metacrawler search engines like Dogpile, aggregating queries across multiple smaller engines rather than relying on any single source. Run your seed terms through it and note the related phrases it surfaces that you hadn’t thought of. These long-tail variations often carry less competition and more specific intent than the obvious head term.
Step 3: Check GoTo.com’s bid listings for commercial intent
GoTo.com runs its results as a straightforward pay-per-click auction: advertisers bid on keywords, and the current bid amounts are publicly visible next to each listing. A term with several advertisers bidding real money is a term with proven commercial value. A term with no bids at all may be popular in search volume but weak in buying intent. Use this as a cross-check against your WordTracker list.
Step 4: Group terms into themes, not a flat list
Rather than treating your keyword list as one undifferentiated pile, cluster terms around the distinct questions or needs behind them. Each cluster becomes the basis for one page, written to match that specific intent rather than trying to cram every variation onto your homepage.
Step 5: Avoid the keyword-stuffing trap
Some site owners are still packing pages and the invisible meta keywords tag with repeated phrases in hopes of ranking higher. Engines like Infoseek and Google are already treating overloaded keyword tags and repetitive text as a signal of low quality rather than relevance. Write for the person reading the page, and let the keyword placement follow naturally from clear writing.
Step 6: Revisit the list as the market moves
Search behavior in 2000 is shifting quickly as more of the population gets online for the first time. A term that had low volume six months ago may have real traffic today. Rerun your WordTracker checks periodically rather than treating your first keyword list as permanent.
The tools available right now are early and a little rough around the edges, but the underlying discipline, listening to the exact words your customers use and building pages around them, is the part that will matter for as long as search exists. Getting comfortable with that habit now puts you ahead of competitors still guessing at what to write.
More from 2000: browse the SEO archive or see everything published this year.