The First 90 Days of SEO for a Roman SME: A Week-by-Week Plan with Semalt

In short: most small businesses in Rome lose their first three months of SEO to work that cannot pay back — redesigns, keyword lists nobody uses, articles nobody asked for. This is the opposite: a week-by-week sequence that fixes what is broken before building anything new, with a checkpoint at each stage so you can tell whether it is working. The measurement side is free to open at semalt.com/authorize.

Ninety days is long enough to fix the structural problems that block everything else, and short enough that you will remember what you changed. It is not long enough to beat an entrenched competitor for the most contested term in the city, and any plan that promises otherwise is selling a date it cannot keep.

Weeks 1–2: find out what is actually true

Nothing gets changed in the first fortnight. The goal is an accurate picture, because almost every business starts with a wrong one.

Count your pages, twice. How many URLs does the site have, and how many are actually indexed? A large gap is the most common finding, and it usually has a boring cause: the same page reachable at several addresses, pages blocked by an old rule nobody remembers, or pages that answer 200 with no content on them.

Read your Search Console, properly. Not the headline graph — the query table. You will usually find that you already rank for things you never targeted, and do not rank for the thing you assumed was your main keyword. Both are useful.

Check the Business Profile. Category, hours, address format, photos, and whether reviews from the last year have replies. For a local business this is often worth more than the entire website.

Write the numbers down. Current impressions, clicks, calls, form submissions, positions for ten queries that matter. Without a baseline, in three months you will have opinions instead of evidence.

Checkpoint: you can state, from data, what your site currently earns from search and where it comes from.

Weeks 3–4: fix the foundations

Now the repairs, in strict order of consequence.

One address per page. Choose whether URLs carry a trailing slash and whether they carry www, then redirect every other form with a 301 and make the canonical agree. This single fix resolves a surprising share of "my pages compete with each other" problems.

One page per service. If your services page lists nine services, it ranks for none of them. Every service that someone searches for separately needs a page of its own, with the words a customer would actually use — not the internal name you use in invoices.

Mobile speed on a real connection. Test on a throttled mobile network, not office fibre. In practice: compress the images, remove the carousel nobody scrolls, and make the address and telephone number visible without interaction.

Titles and descriptions that match intent. A page with high impressions and a low click rate does not have a ranking problem — it has a promise problem, and rewriting the title fixes it within days.

Checkpoint: every service has its own page, every page has one address, and the pages that were already getting impressions now describe themselves accurately.

Weeks 5–8: build what only you can build

Only now does new content make sense, because now it lands on a structure that can hold it.

• Write down the five questions customers asked you this month, in their words, and answer each one on its own page.
• Describe a real job you did — the problem, the constraint, what you chose and why. This outperforms any amount of generic advice.
• Put a named human on the content, with a real role and a verifiable profile. For anything touching money or health, anonymity is a ranking liability as well as a trust one.
• Link the new pages to each other and to the relevant service page. Content that only links back to the homepage wastes the connection.

Checkpoint: five or six pages exist that no competitor could have written, and each is reachable in two clicks from the homepage.

Weeks 9–12: earn signals from outside

Everything so far is under your control. The last month is about the part that is not.

A handful of genuine local mentions. A supplier, a trade association, a neighbourhood publication, a partner you actually work with. Five real mentions are worth more than five hundred directory listings, and directory packages carry risk that real relationships do not.

Checkpoint: reviews are arriving without being chased, and at least three credible sites mention you for a reason unrelated to SEO.

How to read the results honestly

At ninety days, compare against the numbers you wrote down in week two. Expect this shape:

• Local visibility — usually moves first, often within weeks of fixing the Business Profile.
• Long, specific queries — start appearing once the service and question pages exist, typically from week six onward.
• Competitive head terms — mostly unmoved. That is normal and not a failure.
• Contacts — the figure that decides whether any of it worked.

What to keep doing after day 90

The plan above is a repair sequence; it is not a permanent routine. What survives it is short: publish something genuinely specific at a rhythm you can sustain, keep the reviews coming, check the query table monthly rather than the rankings daily, and re-audit the technical basics twice a year — most regressions arrive through a redesign or a plugin, not through an algorithm update.