Link building ROI: the formula, the baseline and what to read at each stage
Link building ROI is (return minus cost) divided by cost, times 100, so a return of twice the cost is 100%. The window decides the figure: one worked example reads −75% at month six and 25% at month twelve for the same campaign, because the cost is spent early while the return builds.

Link building ROI is the return a link campaign produced, less what it cost, divided by what it cost, written as a percentage. The sum is simple; the work is in the inputs: a baseline recorded before the first link goes live, a cost that counts everything the campaign consumed, and a return read from your own Search Console and Google Analytics over a window long enough for the links to have had an effect.
With those inputs, ROI from link building is measurable. This page is a working procedure for how to measure link building ROI, in the order the work happens. It gives the formula and where each input comes from, what to record before the first link goes live, what to read at each stage, two stand-ins for when revenue cannot be attributed, a worked example in which every input is an assumption, and what we report on a placement.
The link building ROI formula and where each input comes from
ROI = (return − cost) ÷ cost × 100
A return of twice the cost is an ROI of 100%, not "2x", and the percentage is kept throughout this page. The percentage lets campaigns of different sizes, and other channels costed on the same basis, sit side by side. Each of the three inputs is taken in turn below.
Return is the extra revenue produced by the pages the links point to, counted over the baseline those pages had before the campaign. It is not all revenue from search. The revenue the pages already produced belongs to the baseline, and only the movement above it belongs to the campaign.
The first place the return is read is the search click. The clicks a page receives from Google Search come from the Performance report in Search Console. You filter the report to the target pages so the rest of the site does not dilute the count. Google's Search Console Help page on the Performance report defines the count: "Clicks: The number of times a user clicked your site from Google Search results."
The second place is what those visits did after they arrived. That is read in Google Analytics 4, where the actions that count are set up as key events. Google's Analytics Help page on key events says: "A key event is an event that measures an action that's particularly important to the success of your business." The return is then one multiplication: extra clicks, multiplied by the share of visits that complete a key event, multiplied by the value of one key event.
One stream sits outside search. Visits that arrive through the placed link itself are referral visits, and they are counted on their own rather than inside the search figure. Tagging them is covered in tagging placed links.
Cost counts every invoice for placements or for a service. It also counts any content written for the campaign. The tool subscriptions used to plan and track the campaign count too. Your own team's hours count as well, at an hourly rate, and leaving them out makes the ROI look better than it is. Use the same cost basis from one campaign to the next, or the two cannot be compared. What placements cost across the market is set out in what link building costs.
Window is the number of months over which the return is counted, and it is stated with the result. The window starts when the first link goes live, not when the order is placed, because nothing can move before a link exists. The same campaign gives a different figure at six months and at twelve, because the cost is spent early while the return builds over time. The worked example below shows both readings of one campaign.
What to record before the first link goes live

Every input on the return side of the formula is a difference: what the page does now against what it did before. Without a dated baseline there is nothing to subtract, so the extra clicks cannot be told apart from the clicks the page already had. Record the following for each target page, on one date, before the first link goes live, and keep the export.
What to record | Where it is read | The maker's own definition |
|---|---|---|
Clicks | Google Search Console, Performance report | Clicks: The number of times a user clicked your site from Google Search results. |
Impressions | Google Search Console, Performance report | Impressions: How many times your site appeared in Search results. |
CTR | Google Search Console, Performance report | CTR (click-through rate): The click count divided by the impression count. |
Average position | Google Search Console, Performance report | Average position: The average position of the topmost result from your site. |
Key events | Google Analytics 4 | A key event is an event that measures an action that's particularly important to the success of your business. |
Referring domains | Ahrefs Site Explorer, Referring domains report | Referring domains are domains from which the target website or web page has one or more backlinks. |
Organic traffic value | Ahrefs Site Explorer, Overview | Organic traffic value is the equivalent monthly cost of traffic from all keywords that the target website/URL ranks for organically, if that traffic was paid via PPC instead. |
The export matters because the report does not keep its history for ever. Google's Analytics Help page on connecting Search Console says: "Search Console keeps data for the last 16 months. As a result, reports in Analytics also include a maximum of 16 months of data." A campaign measured over a year needs its starting figures saved. After the 16 months have passed, the report itself will no longer show them.
Sixteen months feels like a long time on the day the baseline is taken. It stops feeling long when the campaign has run a year and the review lands a few months after that. We've seen people try to rebuild their starting figures from memory, and it doesn't work. What they come back with is roughly what the page used to do, and roughly isn't something you can subtract from. Save the export on the day and it's done.
Each score as its maker defines it is set out in link building metrics. Two tools can give two different counts for the same site's links. The reasons for that are set out in counting backlinks. That difference is the reason the baseline names the tool each figure came from. The later reading has to come from the same tool, or the subtraction compares two measures instead of one.
What to read at each stage

Tracking a link building campaign means reading four stages in order: three leading indicators, then the lagging one, key events and revenue.
First, that the link is live and found. A placed link is first a URL to check. The page is live, and the link sits where it should.
After that, the link appears in a tool when the tool's crawler reaches the page. Ahrefs' help page on how long it takes to find a new backlink, dated 3 September 2026, gives the timing. "So the short answer is that it can take a few days up to some weeks before we crawl your new backlink."
So the referring domains count is the one number in the baseline table that counts the link itself. The date it rises is the tool's crawl date, not the link's publication date. A rise this week can belong to a link placed weeks earlier.
Next, position for the target queries. Average position in the Performance report, for the queries the target page is meant to rank for, is read against the baseline.
The sampled timing figure used here is old and small, and it is quoted with both limits attached. Kristina Kledzik's post on Moz's blog, dated 21 April 2016, describes the sample. "I picked out 76 links pointing to pages which are all similar to each other in content, and we didn't change that content (significantly) for 6 months."
Her finding: "It takes 10 weeks on average to see 1 rank jump". Her own caution: "Keep in mind that I am working with a fairly small data set, so I don't recommend that you promise a 10 spot jump after 22 weeks."
That is 76 links to one company's pages, measured in 2016. Treat it as a scale for patience, not a schedule to promise against. More dated evidence on timing, with each sample's size and method, is in how long link building takes to work.
Then clicks. Clicks follow position, and how many follow a given position depends on the results page for that query, as the next section's click-through figures show. Read clicks to the target pages against the baseline, month by month. Read impressions alongside them, since a page can be shown more often before it is clicked more often.
Last, key events and revenue. Only when clicks have moved is there anything new to count as return. The key events completed by those extra visits are the number the formula uses. A report sent before clicks have moved can still say what has happened: the links are live, the tools have found them, position has or has not moved. What such a report cannot yet do is call any of it a return.
When revenue cannot be attributed: two stand-ins
Some sites cannot tie a sale to a search visit: the sale happens offline, or months later, or through another person. Two stand-ins are in common use for the return side of the formula, and each has a limit stated by its own maker. Neither is revenue, and a report that uses one should say which one it uses.
The first stand-in is organic traffic value in Ahrefs Site Explorer, defined in the baseline table above.
Ahrefs' help page on Traffic value, dated 3 September 2026, says how the figure is made. It says: "We calculate it based on multiplying the monthly organic traffic of each keyword at it's respective ranking position, by their CPC value."
The traffic figure inside that multiplication is itself an estimate. Ahrefs' help page on organic traffic, dated the same day, says so: "NOTE: Remember to treat the organic traffic estimations in Ahrefs as precisely that: estimations."
So a rise in organic traffic value is an estimate built on an estimate. We'd still track it. As a trend line for the same page over time, read against the page's own baseline, it shows whether things are moving, and that's worth having. The trouble starts when it leaves the trend line. Put it in a deck as one number with a currency symbol, and someone upstairs takes it for money the campaign made. Now you're defending a figure that was never revenue. It is not a revenue figure, and a report should not present it as one. We'd keep it as a line that goes up or down, and keep it out of the return column.
The second stand-in is modelled clicks from a position change. Multiply a query's monthly searches by a click-through rate for the new position, and by a rate for the old one, and take the difference.
The sum is easy and the answer comes out looking exact, which is exactly why it gets believed. We'd run it before a campaign starts, on the queries we're deciding between. It gives a rough idea of what a move is worth, and it helps us say where the next placements should go. Where it goes wrong is afterwards, when that difference gets written up as a return. A modelled click was never a visit, and the bigger the search volume, the quicker a small gap between two rates turns into a number nobody can check. It's a planning sum. We'd report it as one.
The rate is the weak input, because a click-through rate is not a fixed property of a position. Advanced Web Ranking's study of organic CTR by position shows how much it moves.
The study is a July 2026 snapshot on a page dated 14 September 2026. It was drawn from more than 1.75 million US results pages a month across desktop and mobile. For its desktop pages, meaning Google US desktop across all types of results page, it reports these average click-through rates:
- position 1: 20.02%
- position 2: 10.36%
- position 3: 3.89%
The same page shows the rate moving from one year to the next. It says: "The average #1 position lost roughly a quarter of its clicks in a year: 20.9% average CTR over the last 12 months versus 28.0% the 12 months before (US, all devices)."
Sille Christensen of AccuRanker, a rank tracking company, wrote on 2 June 2026 in AccuRanker's piece on CTR models: "the CTR for rank 1 can vary from 0% to 75% because the SERP layout can push organic search results up and down the page".
So a forecast built on one fixed rate should say which rate it used, whose it was and from when. Your own Search Console click-through rate for the query is the better input where you have it. This is because it is measured on the results page your page actually appears on.
A worked example, with every input an assumption
The figures below are made up to show the sum; none is a result, a price or a forecast.
- Cost. Six months of placements at $1,000 a month, plus five hours a month of your own time at $40 an hour, is $1,200 a month and $7,200 in all.
- Baseline. The target pages had 400 clicks a month from Google Search before the first link, recorded in Search Console.
- Extra clicks. Over the twelve months from the first link, the same pages had 3,000 more clicks than twelve months at the baseline would have given.
- Return. 2% of visits complete a key event and one key event is worth $150, so 3,000 clicks give 60 key events and $9,000.
- ROI. ($9,000 − $7,200) ÷ $7,200 × 100 = 25%.
Now read the same campaign at month six. The extra clicks so far total 600. Those clicks give 12 key events, and 12 key events give $1,800. The cost counted is the whole $7,200, because by month six it has all been spent. The ROI at that date is −75%.
Nothing about the campaign differs between the two readings except the date. At month six, most of the extra clicks are still to come, while the cost is already counted in full. That one difference is why the window is stated with the figure, and why one reading taken early describes the calendar, not the campaign.
What we report on a placement
You supply the target URL and the anchor text when you order. Before a placement goes out, we check the HTTP status code the target URL returns. After publication, we check that the page is live. We check that the link sits in the body copy, that the anchor text and target URL match the order, and that the page is indexable. Only then is the URL reported to the dashboard. Every live URL is listed there, and the list can be exported as an unbranded file. The full standard is on our quality assurance page.
We use two indexing services as standard for all clients. Between them we see an indexing rate of roughly 95%. Without a service, indexing can take up to 4 weeks. Every delivered link has 12-month replacement cover: if the link or its page is removed in that time, we replace it like for like. On the Managed Service, reporting is monthly and white-label.
What we report is the placement: the live URL and its checks. The return is read in your own Search Console and Analytics, by the steps above.
Frequently asked questions
How do you calculate link building ROI?
Link building ROI is calculated as the return minus the cost, divided by the cost, multiplied by 100, and written as a percentage. The return is the extra revenue the linked pages produced over their recorded baseline. Extra clicks from Google Search are multiplied by the share of visits that complete a key event in Google Analytics 4. This result is then multiplied by the value of one key event. The cost counts every invoice for placements or a service. It also counts content written for the campaign, tool subscriptions and the team's own hours at an hourly rate. A return of twice the cost is an ROI of 100%.
How do you measure a link building campaign's ROI without revenue data?
A link building campaign's ROI can be measured without revenue data by using one of two stand-ins for the return. The first is organic traffic value in Ahrefs, which prices a page's estimated organic traffic at its pay-per-click cost. Its limit is that the traffic figure inside it is an estimate, so the value is an estimate built on an estimate. The second is clicks modelled from a position change, where a query's monthly searches are multiplied by a click-through rate for the new position and for the old one. Its limit is the rate, which moves with the layout of each results page. Advanced Web Ranking's July 2026 data puts the average at 20.02% for position 1 on Google US desktop. Your own Search Console click-through rate for the query is the better input where you have it.
What should I record before a link building campaign starts?
Before a link building campaign starts, record a dated baseline for each target page: clicks, impressions, click-through rate and average position from the Search Console Performance report, key events from Google Analytics 4, and referring domains from the link tool the campaign will be tracked with. Save the export, because Search Console keeps 16 months of data, and a campaign measured over a year needs its starting figures kept. Note which tool each figure came from, so the later reading has one fixed point to subtract.
How long before link building ROI can be measured?
Link building ROI cannot be measured on one fixed schedule, because the measures arrive in stages rather than together. The link is live first; a tool finds it after that, and Ahrefs says its crawler can take a few days up to some weeks to find a new link. Position is read next; the sampled timing figure used here is Moz's 2016 sample of 76 links, which took 10 weeks on average to one rank jump, and its author calls the sample a small data set. Clicks and key events follow position, so the return is counted last, and the window over which it is counted is stated with the figure.
What is a good ROI for link building?
A good ROI for link building is one that compares well with your own other channels, costed on the same basis and read over the same window, because those are figures you can verify. What counts as good for one site depends on the value of one key event. It also depends on the baseline the campaign started from and the window over which the return is counted. There is no sourced benchmark for a link building return, so none is given here.
Stay ahead of the SEO curve
Get the latest link building strategies, SEO tips and industry insights delivered straight to your inbox.


