36% of small French businesses have no website: is that a problem in 2026?

The France Num 2026 barometer, accessed on 17 October 2026, separates three measures: 64% of businesses have a website, 66% have at least one free social media account, and 83% have at least one online visibility solution. These groups overlap. Their percentages cannot be added, and the 36% without websites cannot be treated as businesses with no online presence.

The survey includes 9,655 respondents: 6,786 microbusinesses and 2,869 SMEs, interviewed between 23 March and 18 April 2026. Its results represent the 2024 French microbusiness and SME population by size, sector and region. This is a snapshot of reported practices in the spring, not a continuous technical inventory of every French business domain.

An adoption statistic is not evidence of an individual need. It does not tell you each business’s margins, available capacity or spending budget. An individual investment decision requires an examination of the actual buying journey. The useful starting question is not “Do other companies have a site?” but “What information are our customers looking for, and can they find it today?” National coverage provides context; it does not answer that operational question for an owner.

First, reconstruct a recent customer’s journey: discovery, verification, comparison, contact and decision. For each stage, record the channel used and the question asked. A call about opening hours points to a different issue from a business buyer who cannot forward a clear description of your services to another decision-maker.

A business without a website should pay particular attention to referrals. A recommendation from a partner may lead someone to check the business’s precise activity, service area, identity or previous work. Word of mouth and a website are not necessarily competing channels. A site may support referrals without becoming an independent source of new customers.

For two weeks, record questions that recur before a quotation. Separate observations from assumptions: “Three people asked which towns we serve” is useful evidence; “We lose every online buyer” is not, unless you can substantiate it. Ask new contacts how they found you, but do not make answering that question a condition of receiving help.

Then select one source of friction to reduce. If enquiries repeatedly concern services you do not provide, explaining exclusions may be more valuable than attracting extra visitors. If nobody answers incoming messages, another form simply moves the problem. A website becomes a priority when it supports an identified step and someone can handle what happens next. Start with the customer’s task, rather than with a preferred design or software package.

A business listing mainly makes local information accessible: contact details, hours, directions and activity. Social channels can show current products, work or news and support conversations. A website organises more stable information at an address the business can share directly. None of these channels guarantees visibility or qualified enquiries on its own.

For a shop, first check whether the information needed to plan a visit is consistent. For a service sold through quotations, focus on whether buyers understand the offer and service conditions. This distinction matters more than the number of accounts opened. Even an active social profile may be difficult to navigate when a buyer is looking for one service among older posts.

An owned asset also requires practical control: domain access, recoverable content, clear contractual responsibilities and the ability to change supplier. Having a URL does not remove every technical dependency. Conversely, using a platform may be sensible if its role and limitations are understood and accepted.

Write down a simple division of labour: the listing directs people, social channels demonstrate current activity, and the website explains the offer and prepares contact. Remove unnecessary duplication. Our article on websites for local businesses develops the local journey further; it does not mean that every social presence needs to be replaced or supported by a full redesign.

A shopfront helps identify a place; a business listing and a website serve different functions along the buying journey.

Not launching a site now can make sense if the business operates within a closed contractual network, is already at capacity, or faces a near-term change that would quickly make its content obsolete. These reasons should be checked, however, rather than used to avoid examining the question.

Ask three things: Are you seeking new customers? Can existing customers easily obtain the information they need? Would another investment remove a more urgent bottleneck? A business may need to organise its responses, clarify its services or stabilise operations before paying for another communication channel.

Postponing does not mean forgetting. Record the reason and a trigger for reconsideration: entering a new area, losing an important referral partner, introducing a service or regaining spare capacity. Set a review date too. This protects against two opposite mistakes: buying under pressure and keeping a decision long after the circumstances supporting it have changed.

Even without a website, maintain basic public consistency. Check the trading name, professional phone number, opening hours and channels that someone actually monitors. Do not leave outdated links pointing to abandoned pages. The objective is not to appear everywhere; it is to avoid making an existing relationship unnecessarily difficult. A documented decision to wait is more useful than an unexamined assumption that a site is either essential or pointless.

The minimum scope is not a page count. It is the set of questions a buyer needs answered before making contact. One well-structured page may support a simple offer, while different services or audiences may need separate paths.

Prepare the content before selecting software:

• what you do, who you serve and where you work;
• services you exclude and information needed for a quotation;
• a verifiable business identity and evidence you have permission to publish;
• a clear contact route, an owner and a realistic response timeframe;
• required legal information and an explanation of how personal data is used.

Evidence might include authorised photographs of completed work, a documented process or verifiable qualifications. Do not invent testimonials or customers to fill a template. If previous work cannot be shown, describe your process and limitations clearly rather than presenting a misleading portfolio.

Then test the journey on a phone: understand the service without zooming, find the main action, send an enquiry and receive confirmation. Keep forms limited to necessary information. A location and a short description of the need may be enough to direct an initial conversation; requiring a complete project brief can discourage someone who is still exploring their options.

Finally, ask someone unfamiliar with the business to review the site. Ask what they believe they can buy and what they would do next. Actual misunderstandings are more informative than abstract preferences about colours. Use those observations to adjust the wording before commissioning additional features or more elaborate visual effects.

A clear offer and an explicit next step illustrate the principle; a brochure website does not need a cart or online payments.

The initial quotation is not the entire cost. Compare proposals over the same period, including the domain, hosting, writing, photography, licences, maintenance, content changes and support. Include the internal time needed to supply information, approve copy and respond to enquiries.

There is no credible universal price for this decision: scope, integrations and responsibilities change the work involved. Instead, ask what is included, what is optional and what will be charged after delivery. A lower-cost proposal may be entirely suitable for a narrow requirement. It becomes hard to compare when the work left to the customer is not visible.

Before signing, establish:

• who controls domain and hosting access;
• who fixes an outage and under which conditions;
• who updates software and content;
• how text, images and data can be recovered when the contract ends;
• which recurring charges continue even if no changes are ordered.

The ability to leave is part of the budget. Ask about export formats and transfer limitations, not merely a general promise of ownership. Check content usage rights as well. If you later choose AI-assisted production, our guide to building websites and finding first clients helps distinguish a demonstration from delivery and a sustainable service. Faster production does not remove the need to define responsibilities, future charges or the handover process.

A brochure website is not exempt from obligations because it does not sell online. The French Entreprendre Service Public guidance on mandatory website information, accessed on 17 October 2026, explains requirements according to legal form and activity: identity, contact details, hosting provider and other applicable disclosures. Additional rules depend on factors such as regulated professions, audience and functionality. These references concern the French context, not a universal international checklist.

For a contact form, identify the purpose, lawful basis, recipients, retention period and means of exercising rights. Inform people when their information is collected. The CNIL says its examples of privacy information must be adapted: copying text does not guarantee compliance. An enquiry does not automatically authorise a newsletter subscription.

For trackers, the CNIL distinguishes those requiring consent from those that may be exempt. Do not assume every website needs the same banner or that an analytics tool is automatically exempt. Review the services actually loaded, including embedded content.

Finally, assign technical tasks. Cybermalveillance.gouv.fr’s security recommendations cover updates, restricted access, HTTPS, backups and restoration tests. Ask who performs these tasks and how completion is verified. These measures reduce risk; they do not guarantee that an incident cannot occur. Include them in the operating agreement rather than treating them as an unspecified benefit of buying a website.

Before launch, define a qualified enquiry with the people responsible for responding: a need within the offer, a compatible service area and the possibility of a useful conversation. Qualification does not require knowing a prospect’s budget immediately. Serious projects sometimes begin with an incomplete question.

Keep a simple record: date, self-reported source when known, type of need, response sent, next step and outcome. Analyse out-of-scope requests, unreachable contacts and genuine opportunities separately. A single form-submission total conceals these differences and can make an unhelpful channel look successful.

Then compare periods that make sense for your buying cycle. With low volumes, a change of a few enquiries does not demonstrate the website’s effect. Seasonality, referrals and changes to the offer may also matter. Measure to improve a decision, not to credit every sale to the last page someone visited. Be explicit about what you do not know.

If visitors arrive but enquiries remain unsuitable, review the promise, exclusions and proposed action. If relevant enquiries fail to progress, examine availability, response times and the commercial proposal instead. The website is only one link in the chain. A restrained measurement approach can start with enquiry records and an optional source question, without requiring advertising trackers. Choose additional measurement only when it answers a real decision and its data requirements are understood.

For an agency, the 36% figure is not a list of ready buyers. First establish what “no website identified” means: a domain missing from a listing, a site under another brand, a network page or a confirmed absence. Date public observations and retain their source.

Next, separate the fact from the hypothesis. “Service details are difficult to find” supports a question; “You are losing revenue” asserts an unproven outcome. The owner may know constraints you cannot see: a full schedule, a transition or a deliberately narrow customer base.

Qualification should establish the recognised problem, objective, decision owner, timing and feasible resources. Where these are unknown, suggest a proportionate diagnostic rather than presenting a quotation as self-evidently necessary. The SprintLead page for agencies lets you review its positioning; no prospecting tool replaces that conversation or establishes a buyer’s willingness to invest.

Before contacting anyone, check the rules applicable to the channel, recipient and source of the contact details. Public information is not blanket consent to marketing. Consult the CNIL’s guidance on electronic prospecting, provide appropriate information and respect objections. A refusal, or a channel that already meets the business’s needs, must be allowed to end the approach. Good qualification includes deciding not to pursue an account.

During the first week, document current journeys and incoming questions. Check public contact information and select the main source of friction. Keep the output to one page: observed situation, plausible consequence, missing information and responsible person. Do not order a technical solution yet.

In the second week, write the minimum content and sketch the route to contact. Ask an outsider to test the explanation. If they cannot understand the offer on paper, switching website builders is unlikely to solve the underlying communication problem.

In the third week, compare equivalent scopes: content, access, maintenance, obligations, recurring charges and contract exit. Choose the follow-up measure and the person responsible for replies. Remove features that do not support the selected journey. This makes supplier comparisons more meaningful and keeps future operating work visible.

Finally, in the fourth week, decide whether to launch, postpone or avoid investing for now. Every option should have a reason that can be checked. If launching, start with an end-to-end acceptance test and a review of early enquiries. If postponing, define the trigger for restarting. These thirty days are a suggested decision framework, not a promise of delivery or profitability within a month.