The right keyword is not the most searched one: it is the one typed by someone ready to call you. Targeting « website » is a dead end for a Belgian SME, targeting « build an online shop in Wavre » is a potential order. The whole difference rests on three criteria you can check in one morning: the intent behind the query, the competition already in place, and geographic precision.

The figures set the scene. According to a Backlinko study of 306 million keywords, 91.8 % of Google queries are long tail expressions generating between 1 and 100 searches per month, and the median volume of a keyword is only 10 monthly searches. Ahrefs, for its part, analysed 14 billion pages and found that 96.55 % of them receive no traffic at all from Google. Google also states that around 15 % of daily searches have never been typed before.

For a freelancer or an SME in Belgium, the lesson cuts both ways. The threat: burning six months of content on terms that are far too broad, already held by national players with incomparable budgets. The opportunity: your real market lies in hundreds of small, precise queries, often bilingual and local, that nobody works on seriously.

Why the most searched keyword is rarely the right one

Volume attracts, intent converts

A broad keyword mostly signals curiosity. Someone searching « search optimisation » reads, compares and gathers information. Someone searching « seo agency Walloon Brabant pricing » has already decided to buy and is only looking for who to buy from. The second may bring thirty searches a month against thousands for the first: yet it is the second that fills your diary.

Sort every query into three families: informational (understanding), comparative (choosing between options), transactional (buying or getting in touch). An SME starting out with natural search optimisation concentrates on the last two, then works back up towards the informational once the commercial pages are in place.

The displayed volume is not a promise

Tool estimates are averages, smoothed over twelve months and rounded into brackets. A query announced at « 10 searches per month » may hold sixty in September and zero in July. On a Belgian market, where volumes are already split between French speakers and Dutch speakers, read those figures as an order of magnitude, never as a traffic forecast.

The method in four steps

1. Start from what your customers already say

The best source is not a tool, it is your inbox. Go through your last twenty quote requests and note the exact words used. Add the questions that keep coming up on the phone, and Google's automatic suggestions when you type your trade. In one hour you get thirty to fifty authentic expressions, phrased the way a customer phrases them.

2. Qualify every query

For each expression, ask three questions: can a person typing this become my customer, can I write a genuinely useful page on this subject, and am I credible enough to cover it. Three yes answers: the query enters the plan. A single no: it drops out. That filter usually removes half the list and saves you months of pointless content.

3. Look at page one before you commit

Type the query and observe who occupies the top ten results. If they are directories, marketplaces and national sites with strong authority, move on. If you see companies of a size comparable to yours, or dated articles, the spot is winnable. That single glance is worth more than any theoretical difficulty score, and it rests on the same principles as the basics of natural search optimisation.

4. Assign one main query per page

One main keyword per page, with two or three close variants around it. Two pages targeting the same expression compete with each other, and Google ends up keeping neither.

The Belgian case: two languages, towns and nuances

In Belgium the same intent is searched in two languages and the volumes split accordingly. Translating a French list word for word into Dutch is the costliest reflex: the expressions actually typed in Flanders are different. Each language deserves its own research, as detailed in our guide to the bilingual FR/NL website.

Add the geographic dimension. A query carrying a town name (Brussels, Namur, Antwerp, Waterloo) lowers the volume but multiplies relevance, especially if your catchment area fits within a thirty kilometre radius. Those queries are worked alongside your visibility on Google Maps and local search.

Frequently asked questions

How many keywords should an SME target?

Between fifteen and thirty in the first year, no more. Each main query needs its own well written page. Thirty solid pages are worth more than a hundred weak ones, especially given the share of pages with no traffic at all measured by Ahrefs.

Do you need to pay for a keyword research tool?

Not to get started. Google's suggestions, your customers' questions and a look at the first pages cover most of the need. A paid tool becomes useful beyond fifty or so pages, to track positions over time.

How long does a well chosen keyword take to climb?

Count three to six months for a low competition long tail query, longer on a contested market. A keyword that has not moved after six months signals a query that was too broad, or a page that does not really answer the question asked.

Priority action plan

1. Collect: reread your last twenty incoming requests and note the customers' exact wording.

2. Filter: apply the three qualification questions and drop half the list without regret.

3. Check: open Google's first page for your ten best queries and discard those held by players out of reach.

4. Assign: attach each main query to an existing or new page, and track positions every quarter.

Choosing keywords is an exercise in letting go rather than in ambition. The Belgian SMEs that gain visibility are not the ones aiming highest: they accept aiming small, precise and local, long enough for the sum of all those small queries to become a real acquisition channel.

Unsure which queries deserve priority? We build keyword plans for Belgian SMEs as part of our natural search optimisation support.