Netlinking at agency scale

Netlinking for one client is a pleasant forty-minute task. Pull the Search Console data, pick the pages, write the anchors, place the order. It is the kind of work that feels like SEO.

Netlinking for twenty clients is a full week every month, and it degrades. Not because anyone gets lazy, but because the process has no slack in it.

Where it breaks

The failure is predictable and it always shows up in the same order.

First, the page selection goes. Instead of pulling GSC and computing opportunity, you push whatever page you pushed last time, or the homepage. It is the most time-consuming step and the one whose absence is least visible from outside.

Then the anchors go. Writing five distinct, natural anchors per client per month is a hundred anchors across twenty clients. So you fall back on the platform's "generic" anchor setting, or reuse a phrase, and now every client's profile has a repetition pattern.

Then the history goes. Nobody remembers which URL got a link in April. The spreadsheet exists but is three months stale because updating it is the last step and the first one dropped.

Finally the schedule goes. A client gets missed one month. Nobody notices until the quarterly review, when their link count is 8 instead of 12 and there is no good explanation.

Every one of these is a symptom of the same thing: a per-client process that costs a fixed forty minutes, multiplied by a client count that grew.

What has to become systematic

Three things need to stop being manual before the tenth client, in this order of urgency.

Page selection must come from data, not memory. The opportunity score, which multiplies impressions by missed CTR and a position weighting, is a formula. Formulas do not get tired at client fourteen. Compute it, sort, take the top N. If you are doing this in a spreadsheet with a GSC export you are already most of the way to something reliable.

Anchor variation must be checked against history, not intuition. You cannot remember what you used. Keep a per-URL list of every anchor already spent, and check it before writing the next one. This is a lookup, not a creative act.

The order record must be written at order time. Not at the end of the month, not at reporting time. The moment the order goes out, the row exists: target URL, anchor, host, tier, price, date, and the page's position when you ordered. That last column is what makes the quarterly report possible.

The per-client configuration

The other lever is moving decisions from monthly to once-ever.

Most of what you decide each month for a client does not actually change month to month:

Write the left column down per client, once, and the monthly work shrinks to the right column. A recipe such as two Standard links plus one Premium plus €80 of catalogue budget is a decision you make when you sign the client and revisit twice a year, informed by how you want the budget split.

This sounds trivial and it is the single biggest time saver available, because in practice most agencies re-derive the left column every single month from a Slack thread and a memory.

Guarding against the scale-specific mistakes

Some errors only appear at volume.

Cross-client host collisions. With twenty clients buying on the same marketplaces, the same popular host sites turn up repeatedly. Two clients in the same vertical linked from the same host in the same month is a footprint that neither client can see but a competitor auditing the space can.

Cross-recipe collisions within one client. If a client's monthly plan includes two marketplace lines, both will independently rank the available hosts and both will pick the top one. Deduplicate host domains across every line of an order before confirming, not after.

Subscription drift. Homepage links and similar recurring products renew. Order one every month by habit and the client is paying for twelve overlapping annual subscriptions on the same host. Track which lines are recurring and refresh them yearly, not monthly.

Silent failures. A marketplace returns "not enough sites available in this category" and the order does not go through. If nobody is watching, the client simply gets nothing that month. Whatever your process is, it needs to surface failures loudly rather than logging them somewhere nobody reads.

The realistic end state

The goal is not to remove judgement. It is to spend your judgement on the 20% of decisions that need it.

A well-run monthly cycle at twenty clients looks like this: drafts are prepared automatically for every client at the start of the month, you spend an hour reviewing and adjusting the ones that need it, you confirm, and the delivery tracking runs itself. Which steps can safely be automated is a question worth answering precisely, because automating the wrong one is expensive.

The adjustments are where the value is: the client who just launched a new service page, the one whose competitor moved, the one whose best page is now ranking and should be swapped for the next one down. That is the SEO. Everything before it is logistics, and logistics should not cost you a week a month.