Could a paid link be worth its cost?

A link seller quotes you a price. The domain looks relevant, the page you want to push is on page one, and a few more links would surely move it up. So you buy.

The question that rarely gets asked is whether the move is worth the price. A link costs money once. Whatever it earns arrives slowly, as extra clicks from a better position, and only if the position actually improves. Plenty of purchases never earn back what they cost, and you can often tell before you spend.

This guide sets out the payback calculation, shows how sensitive it is to the two numbers you're least sure of, and walks through a purchase that should be declined.

The payback calculation

Payback is the number of months a purchase takes to earn back its cost:

Payback (months) = link cost ÷ risk-adjusted extra monthly revenue

The bottom half has four parts.

1. Extra clicks at the next realistic position. Not number one. Take the position just above you that a modest set of links could plausibly reach, and work out how many more clicks a month you'd get there than where you are now. Clicks = monthly searches × click-through rate (CTR: the share of searchers who click a result at that position).
2. Value per click. What an average click from this search is worth to you, from your analytics: conversion rate × what a customer is worth. Use first-year value unless you have good evidence of longer retention.
3. Modelled extra revenue = extra clicks × value per click.
4. Confidence. The chance you think the links actually deliver that move. Multiply the modelled revenue by it to get the risk-adjusted figure. Half is a sensible default: it says you think the move is as likely to fail as succeed, which is honest for most link purchases.

For CTR, a typical curve looks like this, though it varies a lot by results page (ads, maps, shopping results and AI answers all take clicks):

If you have Search Console data for the page, your own measured CTR at your current position is better than any curve. The Search Console performance report shows clicks, impressions, CTR and average position for each search.

A worked example

A company sells scheduling software for dental practices. Its /dental-scheduling page ranks at position 9 for "dental scheduling software", which has 880 searches a month. All figures are illustrative.

A link seller offers a package of 8 relevant links at $300 each: $2,400. The pages at 6, 7 and 8 have modestly more referring domains than this page, so position 6 is the realistic target.

Nearly two years to earn back the cost, and that assumes the page reaches position 6 promptly and stays there. This purchase should be declined.

Test how fragile the answer is

Two inputs carry most of the uncertainty: confidence and value per click. Rerun the sum with each one moved.

To check a cell: the $6, 70% case is 16.72 × $6 × 70% = $70.22 a month, and $2,400 ÷ $70.22 = 34.2 months.

Read the table the way you'd read any sensitivity test. The decision is sound if it holds across the plausible range, and fragile if one optimistic input flips it. Here, only one cell drops under 18 months, and that needs you to be 70% sure the links work and your value per click to be right. No cell reaches 12 months. The answer is a clear no.

If the table had shown the opposite pattern, with the default cell under a year and only the pessimistic corners above 18 months, that would be a purchase worth making.

What would have to be true

Turn the calculation round to find the most you should pay:

Maximum link spend = payback threshold (months) × risk-adjusted extra monthly revenue

For this page at the default 50% confidence:

• At a 12-month threshold: 12 × $100.32 = $1,203.84. About four of the eight links.
• At an 18-month threshold: 18 × $100.32 = $1,805.76. Six links, and still short of the package.

Or ask how much demand the purchase needs. To pay back $2,400 in 12 months you need $200 a month risk-adjusted. At $12 a click and 50% confidence, that's 33.3 extra clicks a month, roughly double what this move would bring. The page would need a busier search, a more valuable click, or a cheaper route up.

Declining the purchase isn't the end of the page. Its content might be what's holding it back, and a better page costs far less than eight links. The guide to calculating a page-level backlink gap also helps check whether eight links was ever the right number.

Why 12 and 18 months

A payback threshold is a judgement about how long you trust a gain to last. Three things argue for keeping it short:

• Rankings move. Competitors gain links, results pages change layout, and a position you paid for can slip without you doing anything wrong.
• Links decay. Pages get edited, sites get sold, and links disappear. If you buy links, check they are still in place.
• Money has other uses. A dollar on a slow link purchase is a dollar not spent on content, or on a page that pays back faster.

A workable rule is three bands:

Ship/Scale uses these bands with this calculation: the next realistic position rather than number one, modelled revenue multiplied by a confidence setting that defaults to 50%, and nothing over 12 months bought without a person deciding. The dental page above, at 23.9 months, would not be recommended.

Where this goes wrong

• The CTR curve is a guess for your results page. A results page full of ads, a map pack or an AI answer can halve the clicks at every position. Use your own Search Console data where you have it.
• Value per click is only as good as your attribution. If analytics credits the last click, a comparing search that starts a long buying journey will look less valuable than it is. If you set the figure by hand, write down how you got it.
• The clock starts late. Links take time to be found and counted, so the first months after a purchase often earn nothing. Payback measured from the purchase date is kinder than reality.
• One search understates a page. A page usually ranks for many related searches. Counting only the main one is conservative, and can make you decline a purchase that would pay. If the secondary searches are large and closely related, include them, and be as sceptical about their CTR as about the main one.
• Paid links carry policy risk. Google's spam policies treat links bought to manipulate rankings as link spam, and links Google ignores return nothing. That risk belongs in your confidence figure, and it's one reason confidence should never be close to 100%.
• Confidence is a judgement, not a measurement. It's easy to nudge it up until a purchase you want passes. Set it before you see the answer, and use the same default for every page unless you have a specific reason.

The key points

Work out the extra clicks from the next realistic position, multiply by value per click and by your confidence that the links will work, and divide the link cost by the result. Rerun it with confidence at 30% and 70% and with value per click halved: if the answer only works in the optimistic corner, decline. Buy under 12 months, show 12 to 18 months to a person, and don't recommend anything over 18. When a purchase fails the test, fix the content first or look for a cheaper route up.