SEO in-house, or an agency
What each is good at, and the part nobody can take off your hands.
An agency buys pace and range for a stretch of work. Somebody inside buys continuity and knows the business. Neither one writes your content or approves your changes, and that is where most search work stalls. Settle who owns publishing before you settle who does the work.
Work it out in a minute
Six questions on one page. Names the single thing blocking it, and how long it realistically takes to move.
What we take on
The bottleneck is usually publishing
Most search programs stall because pages do not get written, approved or shipped, not because nobody knows what to do. Whoever you hire, that queue stays yours.
An agency is right for a push
A technical cleanup, a site move, a structure rebuild, a first year of content. Defined work with an end, where breadth of experience across many sites is worth more than knowing yours.
Inside is right for the long run
Somebody who knows the product, the customers and the sales objections writes better pages than a stranger, and they are there next year when the work compounds.
Judge the plan, not the deck
Ask what the first ninety days produce, page by page, and who writes each one. A plan that names the pages is a plan. A plan that names activities is a retainer.
Ask what happens if you stop
The content, the structure and the fixes stay yours. Anything that disappears when the invoice stops was rented rather than built, and that is worth knowing before you sign rather than after.
Keep the accounts
Search Console, analytics, the site itself, and the content in a format you can export. Whoever does the work should be a user in your accounts, not the owner of them.
What each route is actually buying
Be honest about the shape of the work first. A technical cleanup and a structure rebuild is a project with an end. Publishing useful pages every month for two years is not a project, it is a habit, and the two want different arrangements.
An agency is good at the first shape. They have seen the failure modes on fifty sites, they move fast, and they hand back something finished. Paying that rate to grind out a monthly habit for two years is where the value falls away.
Somebody inside is good at the second. They know why the product is priced that way and what the sales team is asked in every call. That knowledge is most of what makes a page worth reading rather than worth skimming.
The part neither one solves is publishing. Pages need writing, somebody has to approve them, and something has to put them live. Where that queue has no owner, both arrangements stall at the same point and both look like the wrong choice.
When you compare proposals, compare what exists at the end of ninety days. Named pages, named fixes, and who writes each one. Proposals that describe activity rather than output are describing a retainer, and the difference is easy to see once you ask.
Whatever you choose, keep Search Console, analytics and the site in your own accounts, and keep the content exportable. That is what makes changing your mind cheap, and changing your mind is a normal thing to do here.
Questions we get
Can one person in-house do all of it?
The writing and the publishing, usually yes. The technical audit, a site move and a structure rebuild are specialist work they may only do once, which is exactly what an outside pair of hands is for.
The common shape is one inside owning the habit, with outside help for the projects.
What should an agency deliver in the first ninety days?
A technical pass with the fixes actually shipped, a structure that reflects what you sell, and a named list of pages with dates and authors. All three should be visible on the site, not only in a report.
If ninety days produces a strategy document, that is a slow start.
How do we tell a good proposal from a bad one?
A good one names pages, fixes and dates, says who writes each thing, and says what you keep if you stop. It is also willing to say which of your ideas will not work.
A proposal built entirely around a monthly activity list is describing time, not outcomes.
Is a freelancer a real option?
Often the best one at small size, particularly for the technical pass and the structure. The risk is the same as any single person: cover, and what happens when they take on a bigger client.
Write down what they do and where it lives, and that risk shrinks a lot.
We tried an agency and nothing happened. Why?
Most often the pages never got written or never got approved. The recommendations existed, the queue behind them did not move, and everybody was busy.
Fix the publishing path first and the same recommendations start working.
What do you do, of these?
The technical pass, the structure, and building the publishing path so pages ship without a bottleneck. We also write, and the arrangement that works best is that somebody inside owns the queue.
The search service page covers what that looks like month by month.
More in the guides and every answer in one place.
Shaheer leads the work, with engineers, writers, filers and analysts behind him. C-suite operations for a San Francisco AI company, Six Sigma on the process side, Anthropic certified on the Model Context Protocol, ten years across eight industries. See what we have built