Anchor Text Optimization: Ratios That Keep a Profile Looking Earned

Anchor advice goes wrong when the percentage is of links, not of referring domains, or when the homepage and the product URL are mashed together. Pick one URL. For this exercise it is https://ledgerline.example/renewals. Ledgerline is a teaching product, not a customer, and the host is a stand-in. Export every referring domain to that URL. One domain that links four times counts once. Classify each domain as brand, naked URL, partial, generic, or exact commercial query. If you cannot classify it in two seconds, it is partial or generic, not exact. Exact means the anchor is the query you are trying to rank, almost unmodified.

The bands below are the guardrails this desk uses when it reads an order. They are not a ranking formula and they are not 'the ratios that get you position one.' A profile inside the bands can still fail because the page is weak or the sites are networks. A profile outside them is simply easier to see as paid. Google's spam policies talk about manipulative linking patterns. A block of identical commercial anchors is that kind of pattern. The bands are how you notice the block before you add to it.

Imagine the renewals URL already has 20 referring domains. This table is a made-up starting profile so the arithmetic is visible. Replace it with your export. Do not add these 20 to a real report.

9+4+3+3+1 is 20. Brand is 45%, inside a 40–60% band. Naked is 20%, inside 15–25%. Partial is 15%, inside 10–20%. Generic is 15%, inside 10–20%. Exact is 1 of 20, which is 5%. On a profile this small that single domain is the whole exact bucket, so the next orders do not add another. Spend the brand and naked buckets, which still have room, or partial if the sentence truly calls for a description. Another 'best contract renewal software' moves exact from 1 of 20 (5%) to 2 of 21 (about 10%) as soon as you add it. One order does that. That is why the percentage is a bad thing to eyeball on a small profile, and why the rule on a young URL is stricter than the table: zero new exact anchors until you have enough brand mentions that one exact match cannot dominate.

Write the anchor in the sheet first. The checkout box will accept whatever you paste, including the same string six times. The assignment below keeps exact at 1. Count the six rows as written: orders 1, 2, 5, and 6 are brand, order 3 is a naked URL, order 4 is partial. Added to the teaching profile that is brand 13, naked 5, partial 4, generic 3, exact 1. The denominator is 26. Brand is 13 of 26 (50%). Naked is 5 of 26 (about 19%). Partial is 4 of 26 (about 15%). Generic stays 3 of 26 (about 12%). Exact stays 1 of 26 (about 4%). Brand is still inside 40–60%, and naked is still inside 15–25%. If your real profile is already heavy on exact, this assignment is the wrong one: you would replace the partial line with another brand and add no partial at all.

Orders 5 and 6 are on the list so you practice deleting them. AI-111 is an applied GenAI magazine at DR 69 and $189. DEV-103 is developer tutorials at DR 71 and $229. A renewal product does not get those rows just because the anchor plan has empty slots. An anchor plan is not a shopping list. If the story does not fit, the slot stays empty and the percentage is computed on the orders you actually place. Four fitted orders are a better profile than six, two of which are off-topic brand mentions. Off-topic brand mentions are how 'natural anchors' still look paid: the words are fine and the site is random.

Brand, which should pass: 'The approval trail is the part spreadsheets lose, which is the problem Ledgerline is built around.' The anchor is the name of the product. A person could have written it.

Naked, which should pass in a docs-like paragraph: 'The field list is documented at ledgerline.example/renewals.' Use this when the brand has already appeared in the same article, or when the referring domain is the third Ledgerline mention that month. It is a boring anchor. Boring is the point.

Partial, which passes only if the words are true: 'Renewal tracking for finance teams is the job the date column was failing at.' The anchor is 'renewal tracking for finance teams,' the same string as order 4. If that is not what the page does, do not use the sentence. Partial match is not a place to hide the exact query by adding two adjectives.

Exact, which fails in almost every draft we see: 'Teams that want the best contract renewal software start by replacing the spreadsheet.' The anchor is doing the marketing. Delete it from order 1 through 6 in the table above. You already have one in the teaching profile. A second one is a choice you should have to justify in the note, in a sentence other than 'it is our keyword.'

Stop. Do not buy a package of generic anchors this week to 'dilute' the exact matches. A burst of guest posts with the anchor 'click here' is a second pattern on top of the first. The next mentions, if you place any, are brand and naked URL, on publications that fit, spaced out. Unlinked mentions of the product name on pages you would want cited anyway also belong in the story, which is the point of the AI search piece, but they do not enter this percentage because they are not links. Keep them in a separate column so you do not pretend they repaired the ratio.

Re-export in 90 days. If exact is still the thick part of the profile, the issue is old anchors, not a shortage of new brand links. Removing or rewriting old paid anchors is a conversation with those publishers, not a shopping trip. This catalog cannot edit someone else's historical article unless that article is the insertion you are buying now.