Estimating Traffic From DR and Keywords: The Truth Gap
Every aftermarket listing leads with the same trio: an authority score, a keyword count, and an implied promise that both translate into visitors. They do not. Authority metrics measure links, keyword counts measure appearances in ranking databases, and the distance between either number and actual traffic is wide enough to lose a bankroll in. This guide maps the gap, explains why it runs in both directions, and shows how to read metric columns without being lied to by them.
What do DR and DA actually measure?
Ahrefs Domain Rating, Moz Domain Authority, Majestic Trust Flow: all are link metrics. They summarize the quantity and quality of pages linking to a domain, and none of them counts a single visitor. They also disagree with each other, by about 26 points on average for the same site across tools, because each vendor crawls a different slice of the web and weighs it differently. Update cadence adds more noise: Ahrefs refreshes DR roughly every 12 hours while Moz DA updates on a roughly monthly cycle, so the same domain can look sharply different depending on which dashboard you opened this week. Useful filters, yes. Traffic evidence, never.
Where do estimated-traffic numbers even come from?
Estimated-visits columns are the output of a pipeline, and every stage adds error. The tool maintains a keyword database with search volumes, observes which domains appear in the top 100 for each query, assigns a click-through assumption to every position, and multiplies its way to a monthly figure. That machinery explains the strange behavior buyers see on expired domains: estimates lag reality because rankings are recrawled on a schedule, dead sites show residual "traffic" for weeks, and two tools disagree because their keyword databases and click curves differ. The estimate is not a lie. It is a model answering the question "what would this visibility be worth if the site were alive", which for an expired domain is precisely the wrong question.
Why can a high-DR domain have zero traffic?
- Links from pages nobody reads. Old blogrolls, directories and footer links pass score-building signals without a human ever clicking through them.
- The ranking content is gone. An expired domain lost its pages during the drop blackout, and rankings belong to pages; a dropped domain ranks for nothing regardless of its score.
- Manufactured profiles. Link campaigns inflate scores deliberately, and a Majestic TF:CF ratio below about 0.3 is the classic signature of citation volume without trust.
- Scores decay slower than visits. Link indexes keep counting a dead site's backlinks for months, so authority is routinely the last metric to notice a domain died.
Why can a low-DR domain have real traffic?
The inverse case trips up buyers who filter too hard. A single citation on a busy Wikipedia article can outdeliver five hundred directory links while barely moving an authority score. Generic names collect type-in visits that involve no links at all. Domains printed on products, in manuals or inside apps receive visitors no crawler can predict. Authority tools are structurally blind to every one of these, which means metric-sorted auction lists systematically underprice them, and disciplined hunters get paid for noticing. The correction is procedural, not clever: check the referral sources and revenue evidence on every candidate whose name suggests habit, whatever the score says.
How should you read the metric columns?
| Listing metric | What it really means | Sanity check |
|---|---|---|
| DR / DA 50+ | Many links of unknown quality | Compare two tools; expect spreads around 26 points; check TF:CF |
| Ranks for 1,200 keywords | Appears somewhere in the top 100 for 1,200 queries | Count top-10 positions; deep rankings earn zero clicks |
| Est. 5,000 visits per month | A model's output, often for a site that no longer exists | Cross-check a second model and confirm the site is even live |
| TF:CF ratio | Trust versus raw citation volume | Below about 0.3 is a red flag for manufactured links |
| Rising keyword chart | A snapshot of history, often pre-expiry | Check the date range against the expiry blackout |
An illustrative two-listing test
Suppose two listings, invented for clarity. Domain A shows DR 54 and 1,800 keywords; a closer look finds three top-10 rankings, the rest parked between positions 40 and 100, and the site dark for two months. Domain B shows DR 18, sixty keywords, and a parking statement earning a few dollars every month for a year. A is a link asset wearing a traffic costume: genuinely useful for an SEO rebuild, worth almost nothing as a visitor stream. B is a small but real traffic asset, because boring recurring cents are the hardest evidence to fake. Neither is a bad buy. The mistake is paying traffic prices for A or link prices for B, and the metric columns alone would have caused both errors. Run the same two-column test, top-10 keywords and documented revenue, on any listing and the costume usually falls off within minutes.
The workflow that closes the gap
Use metrics for what they are good at: compression. Filter the daily flood by authority floors and keyword thresholds in DomCop or the free lists to get a reviewable shortlist. Then switch instruments: check what the traffic tools can actually see, run the verification workflow on anything carrying a traffic premium, and price the domain as what it proved to be: a link asset, a traffic asset, or occasionally both. The truth gap is not a scandal. It is a category error, and buyers who name the asset correctly stop paying for the wrong one.
Frequently asked questions
Is DR or traffic more important when buying an expired domain?
It depends on the play. SEO rebuilds buy links, so DR-class metrics matter most after spam checks. Parking, redirects and audience plays buy visitors, so verified traffic is everything. Decide which asset you are buying before comparing prices.
Why do authority scores differ so much between tools?
Different crawlers, different link indexes, different models, and different refresh cycles: Ahrefs updates roughly every 12 hours, Moz roughly monthly. A spread around 26 points on the same domain is normal, which is why single-tool scores should never settle a price.
Do keyword counts predict traffic?
Weakly at best. Only page-one positions produce meaningful clicks, so a thousand rankings in positions 40-100 round to zero visits. Always ask how many keywords sit in the top 10 before treating a keyword count as demand.
Should I ignore metric columns entirely?
No, use them as what they are: compression for shortlisting, and pricing inputs for link value. The failure mode is not using metrics; it is letting a link metric stand in for visitors nobody verified.
Can a dropped domain keep its rankings?
No. The pages that ranked were deindexed during the blackout, and re-registration does not restore them. Rankings can be re-earned by rebuilding real content, within the boundaries Google set in its March 2024 expired-domain policy.