Let Search Console choose your backlink targets

Most people pick backlink targets by gut feeling. They push the page they want to rank, which is usually the page that makes money, which is usually the page that is hardest to rank.

There is a better method, and the data for it is already sitting in your Google Search Console account.

The pages worth pushing

A backlink moves a page up a few positions. The value of that move depends entirely on where the page starts.

Moving from position 30 to 25 changes nothing, because nobody scrolls that far. Moving from position 3 to 2 is worth something but not much, since you already capture most of the clicks available. The move that pays is page two to page one.

Here is roughly how click-through rate behaves by position on a typical query:

That is why positions 11 to 20 are the sweet spot. The page has already proven it can rank. Google understands the topic and shows the page. It just is not quite strong enough to break through.

Building an opportunity score

Position alone is not enough. A page at position 14 with 30 impressions a month is not worth a €30 link. You need to weigh position against actual demand.

Pull your last 28 days of Search Console data with the page and query dimensions, aggregate by page, and compute:

opportunityScore = impressions × (1 − CTR) × positionBoost

Where positionBoost rewards the positions where a gain pays off most:

position 11-20  → 3.0
position 6-10   → 2.0
position 21-30  → 1.5
everything else → 1.0

Read it as three multiplied signals:

• impressions: how much demand exists. No impressions, no upside.
• (1 − CTR): how much of that demand you are currently missing. A page with 60% CTR has little left to win, while a page with 2% CTR is leaving almost everything on the table.
• positionBoost: how much a few positions are worth from where the page stands.

Then discard anything outside a sane range. Pages above position 3 do not need help. Pages below position 50 have a content or intent problem that no link will fix, and pushing them is throwing money at the wrong diagnosis.

Sort the survivors by score, take the top 20, and normalise to a 0-100 scale so the numbers mean something month to month. That list is your backlink shortlist, ordered.

Turning a page into an anchor

Once you have the page, Search Console gives you the anchor too, indirectly.

For each target page, look at the queries it already ranks for, sorted by impressions. The top query is the topic Google associates with the page. Your anchor text should live in that semantic neighbourhood without being a copy of it.

If a page ranks for "excel training online", good anchors might be:

• "an online Excel course"
• "learn Excel from scratch"
• "this Excel training programme"

Bad anchor: "excel training online". It is the query verbatim, which is exactly what a natural writer would never produce.

The rule of thumb: your anchor should read like something a human wrote inside a sentence. If you can drop it into "For more on this, see ___" and it sounds normal, it works. The full distribution across anchor types is covered in our guide to anchor text strategy.

Avoiding the repetition trap

The failure mode of this method is that it is stable. Search Console data does not change much month to month, so the same three pages will top your list in January, February and March.

Push the same page three months running with three links and you have built an obvious pattern around one URL while the rest of your site got nothing. That concentration is one of the distribution patterns that attract scrutiny.

The fix is a memory. Before generating this month's shortlist, exclude every URL that received a link in the last two months. Drop them from the candidate pool entirely, and only fall back to them if you genuinely do not have enough fresh candidates to fill the campaign.

This does two things at once. It spreads authority across more of your site, and it keeps your link profile from looking machine-generated, because it no longer is.

Making it a monthly routine

The full loop, once a month:

1. Pull 28 days of GSC page and query data
2. Aggregate by page, compute the opportunity score
3. Drop pages outside positions 3-50
4. Drop pages linked in the last two months
5. Take the top N, where N is the number of links your budget buys
6. Read each page's top queries, write one distinct anchor per link
7. Order, record what you did, wait

Step 5 has a subtlety worth knowing: when your budget is split across tiers, pair your biggest spend with your highest-scoring page rather than spreading evenly.

It takes about forty minutes by hand for a single site. For ten client sites it takes a day, which is why most agencies eventually stop doing it properly and start pushing homepages again. If that is where you are heading, automating the mechanical steps is what keeps the method alive at scale.