Finding the right digital agency in Switzerland.

You find the right agency by defining the business goal first, then comparing three providers against the same brief, and finally deciding on the basis of a limited pilot rather than an annual contract. The most common mistake happens before the first conversation: approach a provider without a clear goal and every one of them proposes a different package, after which you are comparing apples with pears.

A sequence that works in practice takes about four weeks:

• Week 1 — the brief. One page of A4: what should happen commercially (more enquiries, a better close rate, lower acquisition cost), by when, within what budget range, who decides internally, which systems already exist. Name the budget range. Withholding it produces quotes that miss what you need.
• Week 2 — the longlist. Five providers from three sources: recommendations from your industry, portfolios of websites you like, and search. Names that surface in AI answers and search results are a legitimate starting point but not proof of quality — both reflect visibility, not results.
• Week 3 — the conversations. Three meetings with the identical brief and the identical questions. Insist that the person who will actually do the work is in the room.
• Week 4 — decision and pilot. Start with a bounded assignment: an audit, one campaign, one page. A pilot costs a few hundred to a few thousand francs and answers the only question a quote leaves open — how these people actually work.

One phone call to a reference in a similar situation beats any case study. Do not ask about the result, that is already in the portfolio. Ask about response times, how mistakes were handled, and how access credentials were transferred.

These eight criteria separate solid providers from well-presented ones. All of them can be checked in a first conversation without any specialist knowledge on your side.

• Depth in the channel that matters to you. An agency that offers everything rarely does everything equally well. Ask which channel is the core and which parts are bought in.
• Who actually does the work. Does one person sell and another deliver? How much senior time sits inside the retainer, and which parts go to subcontractors or offshore teams?
• Measurement competence. Without working conversion tracking, every discussion about success is a matter of opinion. A good agency insists on measuring before optimising.
• Transparency about hours. A retainer is a booking of capacity. Anyone who cannot say how many hours are included and what they are spent on is selling a black box.
• Ownership of your data. Domain, hosting, Google Ads account, GA4, Search Console and the Google Business Profile must be in your name. This is the most important point on the list and the one most often overlooked.
• Comparable references. Not the biggest names, but projects with a similar starting position, a similar budget and a similar market.
• The ability to say no. An agency that has never turned down a job will sell you what is in the brochure rather than what you need.
• Process and availability. Who is your contact, at what rhythm do you get reports, what happens if that person is away, and in which language does the work run?

Deliberately not on the list: agency size, awards and certificates. They correlate poorly with the outcome. Google Partner status, for example, documents ad spend volume and passed exams, not the quality of your campaign.

Ask ten questions, four of them uncomfortable. The uncomfortable ones carry the information: they show whether someone is thinking about your interests or looking for a signature.

• How many hours a month are included, and who delivers them — by role and seniority?
• Which parts of the work go to subcontractors or offshore teams?
• In whose name are the domain, hosting, ad accounts, GA4 and Search Console registered?
• What happens when the contract ends: do I receive access, raw data and everything created, in full?
• Which number sits at the top of your monthly report, and why that one?
• What costs arise besides the fee — licences, tools, media budget?
• How long is the notice period, and what justifies that particular length?
• Uncomfortable: in which case would you advise me against this engagement?
• Uncomfortable: name a project that did not work, and what you changed as a result.
• Uncomfortable: what would have to be true in six months for us both to say it did not work?

The tenth question is the most important. A provider who helps you formulate a stopping criterion is binding themselves to a promise you can check. One who dodges, or points to the complexity of the field, has no such promise. A fourth uncomfortable question is worth asking of providers who argue heavily through rankings: show me a report from a month that went badly, and explain how you responded.

As a documented anchor from a real Swiss price list: local SEO starts at CHF 300 a month, full-service SEO at CHF 800 a month, SEO combined with GEO at CHF 2,500 a month, and a custom-built website sits between CHF 3,000 and CHF 25,000. Those are the prices of DLM Digital, the publisher of this page. They are not meant as an offer here but as a verifiable yardstick against which you can place other quotes.

More important than the absolute figure is how you read one:

• A monthly retainer is capacity, not an outcome. You are buying working hours. Two quotes at the same monthly price can mean very different hour counts — ask before you compare.
• Website prices follow scope. Number of pages, how bespoke the design is, integrations with existing systems, multilingual versions and editorial effort drive the price — not the technology named in the proposal.
• Media budget is not a fee. With Google or Meta Ads, part of the money goes to the platform. Have both line items shown separately, and examine models that bill a percentage of media spend — they reward higher spending, not better results.
• “On request” with no range at all is a warning sign. The final price may well depend on your situation, but every provider knows the order of magnitude after ten minutes.

The market has a hard floor. At Swiss hourly rates, a two-figure monthly budget buys less than one working hour. Packages at that level deliver automated directory listings, not support. At the top end, retainers that consume the margin of several jobs a month are usually oversized for a regional SME. Detailed breakdowns are on SEO costs in Switzerland and what a website costs in Zurich.

The best value for money belongs to the agency that produces the most qualified enquiries per franc invested — and that can only be judged after two to three months, and only with working conversion tracking. Before the work starts, value for money is an estimate; afterwards it is a number. Skip that distinction and you end up buying the lowest hourly rate instead of the best result.

How to make quotes comparable before you lay them side by side:

• Normalise to hours. Monthly price divided by the hours included gives the effective rate. Only that figure is comparable.
• Demand a scope of work. Which tasks are included, which cost extra? A quote without an itemised list can be reinterpreted afterwards at will.
• Count the setup costs. Tracking, migration, initial analysis and content production are often one-off items, and they shift the first-year maths considerably.
• Count your own management time. A cheap provider who needs an hour of your week for clarifications is not cheap.

Start-ups follow their own logic. Their bottleneck is rarely the budget; it is uncertainty about the positioning. So book one channel instead of full service, work with a three-month pilot, and avoid annual contracts while the offer is still changing. An agency that insists on a large starting package while your market is unvalidated is optimising its own utilisation. For SMEs with an established offer the opposite holds: the smallest entry is often the most expensive one, because the foundational work gets stretched across too many months. How ongoing support is structured is described on SEO Agency Zurich.

SEO optimises positions in the classic results list. AEO — Answer Engine Optimization, largely interchangeable with the term GEO — targets the written answer in AI Overviews, Google AI Mode, ChatGPT and Perplexity. Almost any agency can offer both; only the ones that measure both actually command them.

Three checks separate real competence from a new label on an old service:

• Measurement. Does the agency keep a fixed prompt list and log every month whether your company is named and which source was cited? Without that log there is no steering, only assertion.
• Subject of the work. Does it talk about entity consistency, Schema.org markup, answer structure in the opening sentences and source quality — or only about keywords? Language models cite statements that can be checked, not keyword density.
• Honesty about the effect. AI visibility currently produces very few clicks. It pays into mentions, shortlisting and trust. Anyone selling AEO with a traffic promise has not measured the channel.

Pricing rarely separates AEO from SEO cleanly, because the foundations are identical: clean structure, clear answers, consistent company data. A separate AEO engagement therefore only makes sense when AI visibility is business-critical for you and needs intensive measurement. For orientation: at DLM Digital the combination of SEO and GEO starts at CHF 2,500 a month, while the GEO foundations are already included in full-service SEO from CHF 800 a month. What sits behind the term is explained on GEO Agency Switzerland.

A twelve-month SEO engagement is worth it when your website is technically sound, demonstrable search demand exists for what you offer, and the margin on a few extra jobs covers the annual cost. If one of those three is missing, the term is not the problem — the engagement is.

The time horizon is the reason twelve months gets discussed at all. Locally, the first measurable shifts appear in month two to four and a stable position settles between month four and eight; for contested nationwide terms both move out to eight to twelve months. Cancel after three months and you have paid for the build-up phase and given away the harvest. That is precisely why agencies argue through contract length — the reasoning is factually correct, and it doubles as a sales argument.

An annual contract you can sign has four features: quarterly targets stated as numbers rather than as lists of activities, an exit clause after three or six months, reporting that puts enquiries ahead of rankings, and a written rule that access and content stay with you when the contract ends. Without the exit clause, the term mainly protects the provider from the consequences of poor work.

Three situations argue against a long engagement: a relaunch due in a few months — do not optimise twice, postpone the start instead; a purely seasonal or one-off goal, for which paid advertising is the more honest answer; and a lack of internal capacity for approvals, subject-matter copy and access. The cost side of a long engagement is broken down on SEO costs in Switzerland.

You justify an SEO budget internally with a break-even calculation, not with rankings: annual cost divided by the margin on one won enquiry gives the number of extra jobs at which the investment carries itself. A retainer of CHF 800 a month is CHF 9,600 a year; if a new client contributes CHF 5,000 in margin, break-even arrives after two extra jobs. At CHF 200 margin per job, the same point moves to 48 jobs, and the honest consequence is not a bigger budget but a different channel.

The internal one-pager needs six lines:

• Current situation. Today’s numbers as a baseline: organic sessions, enquiries per month, visibility for your ten most important search terms.
• Target metric. One single number to be measured against, usually qualified enquiries per month.
• Measures. Three to five points, not a two-page activity list.
• Cost. Retainer plus one-off setup cost plus your own internal time in hours.
• Break-even. The calculation above, with your margin filled in.
• Exit criterion. What has to be true after six months, otherwise you stop.

The three usual objections in a committee have short answers. “Why not paid advertising?” — ads work immediately, but the visibility is rented and ends with the budget; SEO builds an asset whose cost per enquiry falls over the years. The two are not mutually exclusive. “Why not do it in-house?” — weigh salary, social contributions, tools, onboarding and management time against the retainer; below a permanent half-time role, external is usually cheaper. “What if it does not work?” — that is exactly what line six is for.

Small companies improve their AI search visibility through three levers: one identical company entity everywhere, content that answers a question completely in the first two sentences, and machine-readable structure. None of the three costs budget; all three cost care.

In concrete terms:

• Keep the entity consistent. The same company name, address and phone number on the website, the legal notice, the Google Business Profile, directories and social profiles. Diverging spellings split the entity, and in case of doubt a language model names nobody at all.
• Write answerable content. Question as the heading, the complete answer in the first two sentences, the reasoning afterwards. Concrete figures, prices, deadlines and conditions raise the chance of being cited noticeably, because they can be verified.
• Make the structure machine-readable. One H1, a clean H2 hierarchy, Schema.org markup for the organisation, the services and the FAQ, correct opening hours and location data.
• Build evidence outside your own website. Reviews, industry directories, professional articles, association profiles. Models weight independent sources more heavily than self-description.

You can measure this with a fixed list of ten to twenty prompts that reflect the questions your customers actually ask. Run that list monthly across several systems and log whether your company is named and which source was cited. It is imprecise, but it is the only measurement available — and it reliably shows which of your pages work as a source. Expect mentions rather than traffic: AI answers lead to clicks far less often than classic search results, yet they shape the shortlist before anyone visits a website at all.

An agency is the wrong choice when the need is too small, too continuous or too unclear. In those cases a freelancer, an internal team or deliberately doing nothing is the better decision. This section sits on an agency’s website on purpose, because the alternatives belong in the comparison.

A freelancer is better when you need a single, clearly bounded channel, the budget is small, and you can steer the work yourself. You pay no overhead and speak directly to the person doing the work. The flip side: no cover during illness or holidays, limited breadth of expertise, and a concentration risk if that person changes priorities. For tasks that need technology, content and advertising at the same time, the advantage flips.

An internal team is better when the need permanently occupies more than roughly half a role, the knowledge should stay in the house, and your business is so specialised that external copy always needs expensive reworking. Do the sums honestly: salary, social contributions, workplace, tool licences, onboarding time and management effort. And expect that a single internal person rarely covers technology, content and advertising at a good level simultaneously — many companies do better with one internal owner plus targeted external support than with either model in its pure form.

Doing nothing is better when one of these applies: your offer has no product-market fit yet; there is no demonstrable search demand for your service because your customers buy differently; you have no capacity to serve additional enquiries; or a relaunch is imminent. The most common case, and the one least often said out loud, is a fifth: internally, nobody has the two to four hours a month for approvals, expert input and access. Without that person every engagement fails, however good the agency is. If you are unsure, spend the money on a one-off analysis with a prioritised action plan and decide afterwards.