The intern with a spreadsheet is your PR department

Somebody at your agency is doing PR right now. It's almost certainly not their job.

Unless you started the shop to do press — and some people did, and they're very good at it — outreach is one of those things that gets pushed downhill. To the intern. To the junior who happens to have a light week. To whoever the founder pointed at on Monday and said go get some backlinks for this client.

And it lands on them with no system attached. A spreadsheet, a browser, and good luck.

Big agencies don't work this way. They have people whose entire job is press — who know editors by name, who know which desk covers what, who have been building those relationships for years. That's a headcount line a fifteen-person shop can't carry, and a one-person shop can't even think about.

But the work doesn't go away, because the client is still paying for results that don't happen without it.

Authority comes from other people talking about you. That's been true for twenty years and it's still true — you can fix every heading, ship perfect schema, write the best page in the category, and if nothing anywhere else references the client, the needle doesn't move.

The client isn't going to build it. A few will, the ones who genuinely enjoy this, and they're wonderful. The rest hired you precisely so they'd never have to think about it. It's in the retainer whether or not anyone wrote it down.

So it comes back to whoever's free. And here's what they're actually doing all week: googling. Asking an AI for suggestions. Hunting for listicles that look relevant. Working out who wrote the thing and whether that email address is real. Maybe paying for a subscription or three. Then writing the email — and getting the tone of that email right is its own craft nobody taught them, because a pitch that reads like a link request gets deleted, and a pitch that reads like a person who read the piece gets a reply.

Hours. For one placement, maybe.

Put two published numbers next to each other and the whole economics of this falls apart in your hands.

In a survey of 500 SEO professionals, 76% said they pay $300 or more for a link. 47% pay $500 or more. 16% pay over $1,000. Another analysis puts the range at $150 to $500 including the content.

And the most common monthly SEO retainer is 500 to 1,000 bucks.

So the median retainer buys one link. That's not a great trade-off.

That arithmetic is why link building gets pushed to the cheapest person in the building. It isn't a judgement about how important it is. It's that nobody can afford to have a senior do it.

Worth being clear about the difficulty setting here, because most advice on this subject is written as though everyone's promoting a brand people have heard of.

Getting coverage for a national name is comparatively easy. Editors want them. There's an audience for the story before you write a word.

What you've actually got is a B2B SaaS that launched a new module in March. Nobody's heard of them. There's no news hook, nothing controversial, no founder anybody wants to interview. Your job is to get them named in the same article as three competitors who got in first.

That's a different sport entirely, and it doesn't reward volume. It rewards knowing exactly which piece to go after and exactly why that writer would care.

For a long time this work was measured in links, because a link was the thing you could count. That's shifted, and it changes what you should be chasing.

Practitioners now consistently report that brand mentions and branded search volume correlate more strongly with AI visibility than backlink counts or Domain Rating. Which means a listicle that names your client — even without a link — is doing work that a link used to do, and possibly more of it.

The second thing is concentration. When the assistants answer a question, they pull from a much narrower set of places than Google ever did. Search Engine Land's dataset found Forbes was the only traditional outlet cited across all eleven industries it measured, and that Reddit, Wikipedia, YouTube and Forbes topped every single sector. Conductor's seven-month study found each engine has its own settled habits — ChatGPT anchored on Wikipedia, Perplexity and Gemini leaning heavily on YouTube.

Now, I'd hold the specific percentages loosely. Semrush's analysis of over 100 million citations caught Reddit's share of ChatGPT answers falling from roughly 60% to about 10% inside a single month. These numbers move. The shape is what's stable: a small pool of trusted places, and you're either in it or you aren't.

Which reframes the whole job. It stopped being "get links" and became "get named where the answers are being drawn from." That's not link building. That's PR — the thing nobody at a small agency was hired to do.

And it's the slow one. Insufficient editorial presence in the sources these models learned from is estimated at around 30% of AI-invisibility cases — the single most common cause, and the one that takes longest to fix. Being missing from the structured sources runs three to twelve months. Start now or explain later.

Here's the part I'd build the whole service around, because it turns an open-ended chore into a finite list.

Stop asking where your client could be mentioned. That question has no end and it's what makes this work feel bottomless. Ask a narrower one: where are their competitors already mentioned that they aren't?

That has an answer. It's countable. And every row on it is already qualified — a writer who covered three companies in this category demonstrably covers this category.

Clients pay for that. Not for link building in the abstract, but for being put head to head with the competitor who's already entrenched in the piece. "Your three competitors are in this roundup and you're not" is a sentence that funds work. "We're going to do some link building" is a sentence that gets negotiated down.

Which is where good intelligence earns its money. You get paid, the client gets somewhere they weren't getting on their own — but only if you can put the right list in front of them fast enough to still be selling it this month.

Something people miss entirely when they work this by hand.

Somebody published "the best project management tools of 2025" last December. There's a very good chance they'll write the 2026 edition this December — that's how those pieces work, they get refreshed on an annual cycle because that's what keeps them ranking. Which means there's a window, a few weeks wide, when that writer is actively deciding who's in the new version.

Hit that window and you're helping them do a job they've already scheduled. Miss it by a month and you're a stranger asking for a favour.

So now, on top of everything else: which pieces refresh, when, who's the writer this time, did we already talk to them, what did they say, did they promise anything. Is that another column in the spreadsheet? Another row? One more thing that slips because the person tracking it also has four other jobs and it's Thursday?

This is the point at which the spreadsheet stops being a place to keep information and starts being the reason opportunities get lost.

Which is the quiet killer, because press work is slow and mostly invisible until it isn't.

You pitched eleven writers. Two replied. One said they'd include the client in a piece that runs in March. That's a genuinely good month's work and it looks like nothing on a report that only counts published links. So the line item gets questioned in the budget review, and you lose a retainer you were actually earning.

A commitment is worth reporting. The work happened, the yes exists, the piece just hasn't run yet. Sitting alongside it: here's the target list for next quarter, here's who refreshes when, here's who we're waiting on. Give the client both halves — the work done and the work queued — and the conversation stops being about whether to keep paying you.

Put the outcomes on the invoice, too. A line item with its justification sitting right on it survives a budget review that a line item labelled "SEO" does not.

Press work never finishes. There's always another roundup refreshing, another writer who's moved desks, another competitor getting into a piece you're not in. That's not a flaw — that's retainer work, and almost nobody is billing for it properly.

Now put it against the churn numbers. 40% of clients say they'll switch agency partners within six months, and the two reasons they give are dissatisfaction with delivery and the agency didn't understand our business. A commitment landing every month answers the first. "The writer who covers your category refreshes her roundup in November, and here's why you're in it this year" answers the second, in one sentence.

Which makes this a growth play, not just delivery. Agency leaders rate growing current accounts at 3.6 out of 5 — near enough tied with referrals at the top of every channel.

This is the exact problem I've spent the last stretch building for, so take the enthusiasm with the appropriate pinch of salt.

Coverage Opportunities reads the competitors' coverage and hands back the list — every article that named them, who wrote it, when it ran, whether it looks like it refreshes annually, and the verbatim line where the competitor appears so you can see exactly why the client belongs there. Then the Coverage Pipeline is where you work it: the writers on a board sorted by what you owe them next, the pitch written against the article rather than from a template, sent from your own mailbox, follow-ups on the same thread. It's brand-kit scoped, so twenty clients means twenty separate boards — one menu switch and the contacts, the templates and that client's own case studies come up with it. Nike's press lane never touches the B2B SaaS's. And the report comes out of the same place, in two halves: what's recommended, and what's been done — including the commitments, dated from the day they were given.

But none of the steps change whether you use anything of mine or not. Find where the competitors are and the client isn't. Work out which pieces refresh and when. Write to the writer about their actual article. Track what they said. Report the commitments, not just the links. Plenty of people do all of that by hand and do very well.

One placement a month and everything rests on that one placement. Six or eight and you've got a service — something to price, something to defend at renewal.

The writing was never the bottleneck. Someone still has to send a good email to a real person about a real article, and that part should stay human. It's the hours in front of it — the finding, the qualifying, the tracking — that decide how many you ever get to send.

Go and look at what your best client's competitors are turning up in. Ten minutes, and I'd bet it annoys you.

How much does a backlink actually cost?

In a survey of 500 SEO professionals, 76% pay $300 or more per link, 47% pay $500 or more, and 16% pay over $1,000. Another analysis puts it at $150–$500 including the content. Set that against a median SEO retainer of $500 to $1,000 a month and the problem is obvious.

Do backlinks still matter with AI search?

Yes, but the target moved. Practitioners consistently report that brand mentions and branded search track AI visibility more closely than backlink counts or Domain Rating. A piece that names your client without linking is doing real work now.

Where should a small agency start?

With the competitors. Find the articles that already name them and don't name your client. That list is finite, pre-qualified, and gives you a genuine reason to write — which is worth far more than volume.

What's the best time to pitch a roundup piece?

Shortly before it refreshes. A "best X of 2025" published last December will very likely be rewritten this December, and in that window the writer is actively choosing who's in it. You're helping with a job they've already scheduled rather than asking for a favour.

How do I report press work when nothing's published yet?

Report the commitments. A writer saying yes to a March piece is work completed in January, and it belongs on the January report with the date it was given. Show the target list alongside it so the client can see what's queued.

Can a one-person agency do PR at all?

Yes, but not the way a big agency does it. You won't win on relationships built over a decade. You win by being specific: the right piece, the right writer, the right week, and a reason to write that's visible in the first sentence.

• You're not buying links, you're buying your way out of PR — the economic argument underneath all of this.
• Your client asked ChatGPT about their own industry — why third-party presence is the slow, decisive one.
• Why referrals poison an agency's growth — the same gap-first move, pointed at your own new business.