In-house SEO vs an agency vs nothing: the choice at each stage

Pick the option that buys you progress without stalling the product. The stage decides the mix more than your preferences do.

If cash is tight, bias to DIY and a freelancer per project. If time is tight, bias to an agency scoped to ship pages, not meetings.

Small products win with a simple process you keep. Set it up once, run it every month. Do not add ceremony you will drop.

• Strategy once a year, read monthly: which searches you want, which pages answer them one per topic, how you will earn links and mentions.
• Plan each quarter: the pages to write, the pages to refresh, the technical debts to clear, who does what by when.
• Roadmap across quarters: the plan drawn out so you can trade scope for time without losing the thread.
• KPIs and metrics: pick one KPI you read weekly, like clicks from search or signups from search. Read diagnostics only when it moves.
• Dashboard: Search Console is enough for most small sites. Looker Studio can pull Search Console and GA4 into one view for free.
• Goals: write a number and a date, like two hundred clicks a week from non-brand searches by March.
• Monthly report: what changed in the KPIs and why, what shipped, what next, and the figures as a table. One page.

Example: yourproduct.com writes a strategy that targets “best email parser”, “Zapier alternative”, “/guides/getting-started”, “/integrations/zapier”, and plans a launch on directories for links.

Before you have users, keep SEO light. You want proof there are searches you can win and that these people convert on the product you have.

• What to check: can Google index your site, do you have the core pages, do you have a path to links.
• Why it matters: you avoid building on sand. Indexation and a basic site structure are the floor for any later work.
• How to do it: verify Search Console, submit your sitemap, ship a minimal set of pages: /, /pricing, /features, /guides/getting-started, /blog, /about, /contact.

Spend one hour a week. Read the results pages for your short list of searches. Write one useful page. Fix one technical issue. That is enough now.

Run PPC to sample intent. Bid on five non-brand queries you think will convert. You stop if they do not convert. You keep them if they do, and plan organic pages on them.

• DIY tools: Search Console and Bing Webmaster Tools, PageSpeed Insights, your browser, a crawler’s free tier.
• Skip a hire: an in-house SEO hire or a retained agency is overkill before you have signal.
• Freelancer edge: buy a small audit or a strategy if you want judgement once without a retainer.

A fixed page looks like this: “/guides/email-parser-setup” that answers the job, shows the product, and links to “/integrations/zapier” and “/pricing”.

Once you see impressions and some clicks, raise the tempo. Keep control as founder. Bring in a freelancer to ship pages and fixes faster than you can alone.

• What to check: which queries already show you, which pages get impressions, which are missing from the plan.
• Why it matters: you double down on what Search Console tells you is working. You avoid writing ten posts no one sees.
• How to do it: pull the Performance report, filter non-brand queries, list striking distance queries in positions 4 to 15, and brief pages to win them.

1. 1

   Write and ship core topic pages

   Create one page per topic: category, best-of lists, comparisons, jobs to be done, integrations, and brand. Example: “/alternatives/zapier” and “/compare/zapier-vs-yourproduct”.

2. 2

   Refresh what almost ranks

   If “/blog/email-parsing-rules” sits at position 8, add examples, FAQs, and internal links from “/features/rules-engine”. Aim to move to the top three.

3. 3

   Clear small technical debts

   Fix title tags, meta descriptions, broken links, and indexation issues. Keep Core Web Vitals in the green: about 2.5 s LCP, 0.1 CLS, 200 ms INP.

4. 4

   Earn simple links and mentions

   Ship a small tool or data post, list yourself in relevant directories, and include your product in community roundups you already use.

Hire a freelancer by scope. Example: three pages briefed and written, a technical sweep of the site, and one day a month on updates. You stay the editor and owner of the plan.

When search becomes a channel, bring continuity inside. An in-house lead keeps the strategy, plan, roadmap, goals and reporting in one head. Flex the rest with an agency or freelancers.

• What to check: is search now a top source of signups or revenue, and do you have enough work to keep someone busy each week.
• Why it matters: you avoid a full-time hire too soon, and you avoid bouncing between agencies with no memory.
• How to do it: scope the next two quarters. If you have ten to twenty pages to produce, a content refresh batch, technical work, and link projects, hire a lead and contract execution.

Use the agency for production runs. Example: a twenty page alternatives and comparisons family, shipped to your briefs, with design, QA and internal links mapped from “/features/*” and “/pricing”.

Keep sensitive work in-house. Strategy, budgets, analytics, conversion tracking, and anything that touches your positioning or roadmap should live with you or your lead.

Buy the thing they are built to sell. Do not pay an agency for judgement if you only need three pages written. Do not hire full-time for a once-a-year audit.

Costs come from work, not titles. Agencies use monthly retainers. Freelancers bill hourly, daily or per project. Tools bill monthly. Price ranges online vary by country, scope and the seller’s marketing. Size your budget by the work and by what a search click is worth to you, not by a claimed average.

Pick one KPI. Make it the number you read weekly. Make everything else support it only when it moves. Keep the dashboard simple enough you open it.

• KPI examples: clicks from search, signups from search, revenue from search.
• Diagnostics: impressions, average position, indexed pages, Core Web Vitals, referring domains. Look only when the KPI shifts.
• Dashboard: Search Console is enough for most small sites. Looker Studio can join Search Console and GA4 for free.
• Goal format: a figure and a date, never a rank. Example: two hundred weekly non-brand clicks by March.
• Report format: one page each month, with what changed in the KPIs and why, what shipped, what next, and a table of the figures.

Most risk is in opportunity cost. Months on the wrong pages or a broken site hurt more than almost any bill. Avoid obvious traps early.

• Promises to guarantee rankings or first positions. No one can.
• Secret techniques. Real work is obvious once you see it.
• Links for sale. Buying links that pass ranking violates Google’s spam policies.
• Scaled content abuse. Do not produce many pages to manipulate rankings rather than help people, however you make them.
• Site reputation abuse and expired domain abuse. You do not need tricks. You need pages with value of their own.
• AI-assisted writing is fine. Pages made at scale without value are not. Fix the page family by giving each page data or an answer worth reading.

Start from the work, not a line in a blog post about “average SEO costs”. Write what you need to ship this quarter and next, then price that work in the market you have access to.

1. 1

   List the work by quarter

   Example: Q1, six pages, a refresh pass on four posts, fix Core Web Vitals on templates, and five listings to pitch. Q2, ten comparison pages and internal linking.

2. 2

   Attach value by query

   If “email parser for invoices” sends signups in PPC, a page for it is worth more than a generic post. Use the revenue per signup to bound the value of the clicks.

3. 3

   Decide who does what

   Founder edits briefs and owns Search Console. Freelancer writes pages and does a technical sweep. Agency ships a batch of comparisons. In-house lead sets goals and reports.

4. 4

   Pick the tool spend you use weekly

   Buy what you open every week. Keep the rest to free tiers. Search Console covers indexation and performance. PageSpeed Insights covers speed debugging.

A fixed outcome looks like this: twelve pages live, Core Web Vitals improved on product and blog templates, five listings secured with branded mentions, and internal links from “/features/*” to each new page.

Most small products end with a hybrid. The founder or an in-house lead owns the strategy and KPIs. External help ships pages, fixes, and links in runs. Keep scope tight and visible.

• Strategy cadence: write it once a year and read it monthly. Invite your agency or freelancer to the read, not to rewrite it each time.
• Briefs first: you own search intent, page purpose, and acceptance criteria. External writers fill the page, not the plan.
• QA and analytics in-house: you own URL structure, internal links, and tracking of signups from search.
• SEM and SEO together: run PPC to learn which queries convert, build organic pages to hold them. Paid stops when the budget stops. Organic keeps sending clicks after the work.

If you want judgement once without a retainer, a one-off report is an honest option. Porteur is one report bought once, 99 € or 199 € with product analytics, and you run the plan yourself after that.