20-person consultancy

From 10 to 150 visitors a day: rebuilding a consultancy's inbound engine

Positioning audited, a content strategy designed with generative AI, the site rebuilt, a CRM installed, then an assistant wired into every measurement source.

Client
20-person consultancy
Secteur
Geomarketing and site selection
Période
2026
Technologies
ClaudeHubSpotVercelGoogle Search ConsoleGoogle AnalyticsGoogle Ads

×15

daily traffic, from 10 to 150 visitors a day

Inbound enquiries up from one every three days to two a day

One inbound enquiry in four turns into signed work

This consultancy designs retail site selection studies for chains and for people opening their first outlet. The expertise is real and long established. The visibility was not: ten visitors a day on the website, one inbound enquiry every three days, and a business running almost entirely on referrals.

The firm has around twenty people. It has no marketing team and was not about to hire one. So the question was the one most owners of small and mid-sized companies end up asking. How do you build a steady flow of inbound enquiries without hiring the function that normally comes with it.

Start with positioning, not with content

The temptation in this situation is to publish. Write articles, post on social media, wait. That is why so many company blogs return nothing. They answer the questions the company asks itself, not the ones its buyers type into a search box.

So the work started with a positioning audit. What does the firm actually sell, to whom, and how is it different from its competitors. The analysis was done with generative artificial intelligence as a working partner: the real service line set against the search demand in the sector, against what competitors already cover, and against the subjects nobody was answering properly.

What came out was a map of topics ranked by search volume and difficulty, and, more useful still, a list of open ground. That is where most of the gain came from. The traffic competitors were leaving on the table could be identified. It only had to be gone after.

The site rebuilt in a few weeks

The old site could not carry that strategy. Its structure had no room for topic pages, and the path to getting in touch was long.

It was rebuilt quickly, with generative AI helping to write the code. That way of working changes how a company treats its website. A change that used to be argued over in a meeting and scheduled for the following quarter now happens the same day. The site stops being a finished deliverable and becomes something you adjust in response to the numbers.

HubSpot went in at the same time, so that enquiries land somewhere and so that the firm can finally see what each source is actually worth.

The step that changes everything: an assistant that sees what you see

This is the most important part of the engagement and the hardest to explain.

Once the setup was in place, the assistant was connected to every source: the CRM, the site’s source code, the hosting statistics, the search console, the ad account and the audience measurement tool. Technically this runs over a recent protocol that lets an assistant read outside tools rather than making do with whatever someone pastes into a conversation.

The practical consequence is large. As long as an assistant works from a description, it produces general advice, the kind you read in any marketing article. As soon as it can see at the same time the queries bringing traffic, the pages that convert, the contacts created in the CRM and the code of the site, it stops recommending and starts pointing. Here is the page that is missing. Here is the query where the site ranks tenth with no content dedicated to it. Here is the change to make.

The working rhythm becomes monthly and short. Look at what moved, decide on two or three actions, apply them the same day.

The results

Within a few months daily traffic went from ten visitors to 150. Inbound enquiries went from one every three days to two a day. More to the point, quality followed: one inbound enquiry in four turns into signed work, which is only possible because the content attracts people whose need genuinely matches the offer.

The whole setup now takes a few hours a month to maintain. For a company this size that is the real test of success. Not how well a launch performed, but whether the pace can be held without hiring anyone.

What transfers

Three things hold for almost any company of this size.

First, the gain comes from analysis before it comes from production. Knowing what to write is worth more than being able to write quickly.

Second, measure before you publish. Without a CRM and without measurement tools connected, you cannot know what worked, so you cannot do it again.

Third, the decisive difference is access to data. An assistant that can see your figures and your site is working on your business. An assistant that sees nothing produces advice that would fit anybody.

The end client is not named. The figures quoted are those of the engagement.

Next step

Where would you start?

Ten days to map where your teams lose time, put a figure on each lever and name the first system to build. If AI is not the right answer, I will tell you so.

Diagnostic

Étape 1

10 days

€9,500 excl. VAT · fixed fee, one area of the business

  • A map of where the time actually goes in the area audited
  • Each lever costed in hours freed, ranked by effort and by gain
  • The data note: what leaves your systems, and where it is processed
  • A three-month plan, naming the first system to build