1. Sell: getting out of order-by-message

Selling through messages works up to about thirty orders a month. Beyond that the human cost explodes: negotiating availability, confirming addresses, typing errors, forgotten replies. A properly designed checkout absorbs that volume with no extra staff, takes payment at night and on Sundays, and produces usable data: most-viewed products, the step where buyers drop off, real average basket.

• Permanent availability : nearly 40% of online orders in Tunisia are placed after 8 p.m.
• Zero negotiation: the price is displayed, so is availability, and the order completes itself.
• Traceability: every order has a number, a status and an invoice, essential as soon as you grow.

2. Gain autonomy: take back the margin and the data

On a marketplace you pay a commission, you are compared with competitors on the same page, and you recover neither the email address nor the purchase history of your buyers. On your own site the margin stays with you, the prices are yours, and above all the customer base belongs to you. That base is what generates revenue six months later without any advertising budget.

One simple metric to watch: the share of revenue coming from customers you already know. Below 20%, you are buying every sale. Above 40%, your store has become an asset.

3. Build loyalty: the cheapest sale is the second one

1. Send a genuinely useful follow-up email after delivery (care instructions, how-to) not a sales push.
2. Offer one-click reordering for consumables: it is the single most profitable lever in a store.
3. Segment your list by purchased category instead of sending the same offer to everyone.
4. Handle returns fast and without argument: a well-managed return builds more loyalty than a frictionless sale.
5. Reward tenure, not just acquisition: welcome codes attract buyers who never come back.

The technical conditions for success

• Mobile speed under three seconds: every extra second costs conversion points.
• Short checkout, purchase possible without an account, shipping costs shown on the product page.
• Complete product sheets: dimensions, material, warranty, real delivery time per governorate.
• Mixed payment (online and on delivery) so you lose no category of buyer.
• Clean analytics: without funnel measurement, every optimisation is guesswork.

An e-commerce site is not profitable in month one, and never by magic. It becomes profitable when three functions chain together: a sale that happens without you, a margin you keep, and a customer who returns. Everything else (the theme, the colours, the number of products) is just the visible surface of that mechanism.

The six components of a working store

1. The catalogue. Product sheets with variants (size, colour), multiple photos, live availability and prices : VAT included or not, depending on your tax regime.
2. Search and filters. Beyond thirty products, a customer who cannot find an item in two clicks leaves the site.
3. Cart and checkout. The most sensitive point: every unnecessary field loses orders. A guest must be able to buy without creating an account.
4. Payment. Online, on delivery, or both. In Tunisia cash on delivery remains dominant and must be treated as a primary method, not an option.
5. Logistics. Delivery fees per governorate, announced lead times, tracking numbers, returns handling.
6. The back office. Orders, stock, customers, invoices, accounting exports: this is what you will use every single day.

Online payment in Tunisia, concretely

Three routes coexist. The Clictopay bank gateway, tied to your bank, requires a merchant file and several weeks of administrative delay. Providers such as Paymee or Flouci integrate faster, with a per-transaction commission. Finally, cash on delivery needs no integration but generates a significant cancellation rate that must be priced into your margins. Most profitable Tunisian stores combine all three.

Start the online payment application on day one of the project. It is the longest administrative step, and the one that most often delays a launch that is otherwise technically ready.

Own store or marketplace?

Selling on a marketplace brings immediate traffic but imposes a commission, direct competition on the same page and no usable customer data. Your own store requires building an audience, but keeps the margin, the relationship and pricing freedom. The most robust strategy uses the marketplace as an acquisition channel and the store as a retention channel.

What actually makes a store sell

• Real photographs, plenty of them, with a sense of scale: the single biggest conversion factor, far ahead of design.
• Shipping costs shown early, never revealed at the last checkout step.
• Visible customer reviews, including average ones: their total absence worries buyers more than a 4 out of 5.
• A product page that answers objections: material, dimensions, warranty, returns, real lead time.
• Mobile load time under three seconds, a minimum condition on the Tunisian network.

An online store is not a website with a buy button. It is a complete sales tool whose other half (logistics, stock, customer service) happens off screen. And it is precisely that half which decides whether it is profitable.

How a DIY platform actually works

You are not renting software, you are renting a full service: hosting, security, updates, visual editor, templates and support, billed monthly. In exchange you have no access to the source code, you do not choose your server, and you cannot migrate the site as it is. That is a perfectly acceptable contract : as long as you know it before signing.

What DIY does very well

• Going live fast. A presentable site in a few days, with no intermediary and no requirements document.
• Testing an idea. To validate an offer before investing, it is the most rational tool available.
• Controlling entry cost. From 30 to 120 TND per month, with no upfront cash.
• Avoiding maintenance. Backups, certificates, updates: the platform handles all of it.

Where the DIY system hits its limits

• Performance. Visual editors generate heavy code. On mobile data the speed gap with a developed site is immediately visible.
• Advanced SEO. Constrained URL structure, limited tag control, often artificial multilingual handling : a serious handicap for a trilingual French, English and Arabic site.
• Lock-in. You cannot export your site. Changing platform means rebuilding everything, and starting over in search if the URLs are not preserved.
• Long-run cost. 90 TND a month is 5,400 TND over five years, without owning the asset at the end.
• Business integrations. Connecting an ERP, a point of sale or a local delivery API quickly becomes impossible or very expensive.

Before choosing a DIY platform, ask a single question: can I export my content and my URLs tomorrow? If the answer is no, you are not building an asset, you are renting a shop window.

The exact moment to move on

1. Organic search becomes your primary acquisition channel and you hit the platform’s technical ceiling.
2. You pass a few hundred orders a month and transaction fees become significant.
3. You need to connect the site to an internal tool : stock, accounting, logistics.
4. You need genuine multilingual support with clean per-language URLs and correct alternate tags.
5. You lose time every week working around the editor: the hidden cost already exceeds the price of a build.

Our position as an agency is unambiguous: DIY is an excellent starting point and a poor destination. Start there if your offer is not yet settled. Leave as soon as the site becomes a revenue channel : and on that day, insist on keeping your URLs.

The price ranges actually charged in Tunisia

These figures reflect projects delivered by agencies and established freelancers in Tunis, Sfax or Sousse, excluding promotional stunts and "free website" offers funded by a long subscription. They cover design, integration, development, deployment and admin training; they do not include copywriting, translation, photography or advertising budget. Amounts are quoted before tax : 19% VAT is added by a registered provider. Euro equivalents are indicative, based on 1 € ≈ 3.3 TND.

• Brochure site, 5 to 8 pages : 1,200 to 3,500 TND (€360 to €1,060). Customised professional theme, contact form, basic SEO, a single language. The typical format for craftspeople, small practices and young companies.
• Bespoke multilingual brochure site : 3,500 to 9,000 TND (€1,060 to €2,730). Original art direction, two or three languages, blog, technical SEO, animation, analytics.
• Product catalogue without online payment : 4,000 to 12,000 TND (€1,210 to €3,640). Product sheets, search, filters, quote requests, initial catalogue import.
• E-commerce store : 7,000 to 25,000 TND (€2,120 to €7,580). Online payment (Clictopay, Paymee, Flouci), stock management, carriers, customer accounts, invoicing.
• Bespoke platform or business application : 25,000 to 90,000 TND (€7,580 to €27,300). Custom development, APIs, user areas, dedicated back office, formal acceptance testing.
• Redesign of an existing site : 2,500 to 15,000 TND. The price depends mostly on content migration, the redirect plan and the volume of indexed URLs to preserve.

A quote under 900 TND for a "complete" website almost always hides a theme resold as is, with no written content, no optimisation and no maintenance. The real cost resurfaces six to twelve months later, when everything has to be rebuilt.

The eight factors that genuinely move the bill

For an apparently identical scope, two quotes can differ threefold. In nearly every case the gap comes from these eight variables : the ones to have spelled out in writing before you compare anything. Our advice: walk this list point by point with every provider you consult. You will usually discover that the cheapest quote does not cover three or four of these lines, and that the real gap between proposals is far smaller than it looked.

• The number of templates, not pages. Ten pages built on one layout cost less than four pages with four different layouts.
• Bespoke design or adapted theme. Original art direction typically accounts for 30% to 40% of a brochure site budget.
• The number of languages. Each language adds translation, typographic adaptation (Arabic requires a mirrored layout) and its own internal linking for SEO.
• The e-commerce funnel. Payment gateway, delivery, VAT, stock, returns: the purchase path often demands more work than the rest of the site combined.
• Integrations. ERP, point of sale, CRM, marketplace, carrier APIs: every connection is priced separately and must be tested.
• Content. Copywriting, translation, photography and video are rarely included. Budget 60 to 250 TND per written and optimised page.
• Search optimisation. URL structure, structured data, performance, keyword research: a site optimised from the start costs 15% to 25% more and returns considerably more.
• The deadline. Delivering in three weeks instead of eight mobilises several people in parallel, and is invoiced accordingly.

The recurring budget nobody mentions

• Domain name : 30 to 90 TND per year for a .tn, 40 to 70 TND for a .com.
• Hosting : 150 to 900 TND per year for a brochure site or small shop; more as traffic or catalogue grows.
• Maintenance, backups and updates : 400 to 2,400 TND per year. This is the line that prevents a white screen on a Friday night and a hack a month later.
• Transaction fees : 1% to 3% of collected revenue depending on the gateway, on an e-commerce site.
• Content and acquisition : articles, visuals, advertising. This budget, not the website itself, decides the results after launch.

How to read a web agency quote

1. Confirm that ownership of the code, domain and credentials transfers to you on delivery, with no lock-in clause.
2. Ask whether the site is manageable by your own team, with which tool and what training.
3. Insist on a line-by-line breakdown: design, integration, development, content, SEO, training, maintenance.
4. Check the number of revision rounds included in the design phase: it is the main source of overrun.
5. Look at what happens after launch: warranty period, included support time, response time when something breaks.
6. Compare like for like. A 2,000 TND quote with no content and no SEO is not cheaper than a 4,500 TND quote that includes both : it is incomplete.

Frequently asked questions about website pricing

• Can I get a website for 500 TND? Yes, with a website builder and your own copywriting effort. No, if you expect guidance, original content and a site capable of ranking on Google.
• Do I have to pay a monthly subscription? No. Hosting and maintenance can be annual and optional. Be wary of offers where cancelling the subscription makes the site disappear: you never owned it.
• How long until launch? Three to five weeks for a brochure site, six to twelve weeks for a shop : provided copy and photos arrive on time. Content is the number one cause of delay.
• Is it paid in one instalment? Rarely. The most common rhythm is 40% on order, 40% on design approval, 20% on delivery.
• Does a more expensive site earn more? Not mechanically. What earns is the fit between the site, your sales cycle and the acquisition budget behind it. A 15,000 TND site with no content budget produces less than a 5,000 TND site fed every month.

How we price projects at WithNet

We start from the commercial objective, never from the page count. How many enquiries a month are you aiming for? What is your average order value? What is a customer worth over three years? A concrete example: a service business whose average client is worth 4,000 TND over three years needs only three new clients a year to pay back a 6,000 TND website : a realistic target in year one with local search properly handled. Conversely, a shop with a 45 TND average basket has to think in volume, therefore in automation and cost per acquisition, not in aesthetics. Those three numbers define a sensible budget far better than a feature list. Only then do we split the project into deliverable phases: a foundation that works and earns first, extensions later, funded by the results of the foundation. It is the only method we know that keeps a website an investment rather than a cost line.

Two different moments of the customer relationship

The website serves discovery: it is indexed, reachable in one click, requires no installation, and it captures search demand. The mobile app serves loyalty: it addresses someone who already knows you, who agrees to give you space on their home screen, and who will come back often. Nobody installs the app of a company they have just discovered. That asymmetry decides the build order.

What the two share technically

• The API and the database. Catalogue, accounts, orders and stock must live in one place, cleanly exposed to both clients.
• Authentication. An account created on the website must work in the app, with no second sign-up.
• The design system. Colours, typography, icons, components: a shared reference prevents two diverging visual identities.
• Editorial content. The same CMS can feed website pages and the news screens of the app.
• Deep links. A website URL should open the matching screen in the app when it is installed.

The real shared deliverable is the API. If your provider builds the app on its own database, separate from the site, you have just bought two systems to maintain : and two invoices for every future change.

When an app is genuinely justified

1. Your customers return at least weekly: recurring orders, account tracking, bookings, deliveries.
2. You need accepted push notifications, the camera, continuous geolocation or offline operation.
3. Your mobile order volume already runs into several hundred per month through the website.
4. You have field logistics or after-sales operations to equip, which justifies an internal app.

The alternative you should not ignore: the PWA

A progressive web app installs straight from the browser, displays full screen, works partly offline and sends notifications on Android. It costs a fraction of a dual-platform native build and depends on no app store. For most Tunisian retailers and service businesses it is the most profitable intermediate step before considering native development.

The build order we recommend

Website first, with a clean API from day one. Measure mobile behaviour for six months. Then a PWA if recurring usage is confirmed. Native app last, once the numbers (visit frequency, return rate, customer value) justify the budget and the annual maintenance that comes with it. Any other sequence means paying for infrastructure before proving it will be used.

Search is becoming conversational

People no longer type "website price"; they type "how much does a website cost for a small company in Tunis". AI-assisted engines answer more and more directly, citing pages that treat one precise question completely. The practical consequence: a generic "Our services" page loses ground to a page that answers a single question with figures, cases and a clear conclusion.

Short video became the default shop window

• Vertical format, 15 to 40 seconds, subtitled: most views happen with the sound off.
• Shown, not told. A finished job, a before-and-after, a technical gesture beats any brand speech.
• Published everywhere. The same file feeds Reels, TikTok, YouTube Shorts and your homepage.
• Measured on watch time, not view count: three extra seconds of retention are worth a thousand impressions.

First-party data is back on top

As third-party cookies fade and privacy rules tighten, ad targeting keeps losing precision. The companies coping best are those that built their own base: email addresses, WhatsApp numbers, purchase history. A genuinely opted-in list of 2,000 customers is now worth more than an equivalent ad budget, because it depends on no platform.

The best marketing investment for a Tunisian SME in 2025 is also the oldest one: collect your customers’ contact details properly, and write them something useful once a month.

Local SEO still has the best cost-to-result ratio

A complete Google Business profile, recent reviews, accurate opening hours and real photographs generate, for most retailers and service providers, more enquiries than a paid campaign. It is unspectacular, entirely free, and yet the majority of profiles we audit in Tunis are incomplete or out of date.

Trends you can safely ignore

1. Mass AI-generated content published without editing: it is recognisable, it does not rank, and it damages the brand.
2. Opening an account on every new network: two channels properly maintained beat a ghost presence on six.
3. Annual logo redesigns: brand recognition is built by repetition, not by novelty.
4. Vanity metrics (followers, impressions, reach) as soon as they connect to no sale at all.

The constant behind all these trends does not change: answer precisely a question a customer is asking, at the moment they ask it, on a medium you own. Everything else is formatting.

What a web design agency does

It designs and builds the object. The work starts with understanding your business and ends with a live website: site structure, art direction, mock-ups, integration, development, content, deployment. It is judged on the quality of what is delivered (clarity of the journey, performance, brand fidelity, technical compliance) and on hitting deadlines.

• Typical deliverables: requirements document, mock-ups, working site, admin training, documentation.
• Billing: per project, usually in three instalments, with an optional annual maintenance contract.
• Horizon: a few weeks to a few months, with a clear beginning and end.

What a web marketing agency does

It brings traffic to the object and turns it into customers. Organic search, paid advertising, social media, email, data analysis, conversion testing: its territory starts where the designer’s ends. It is judged on metrics (cost per lead, conversion rate, return on ad spend) not on a graphic deliverable.

• Typical deliverables: acquisition plan, campaigns, editorial calendar, monthly reports, optimisation recommendations.
• Billing: recurring monthly retainer, often plus a percentage of media spend.
• Horizon: continuous. First organic results rarely appear before three to six months.

Simple rule: web design produces an asset, web marketing produces a flow. One is owned, the other is maintained. Neither replaces the other.

Where the two trades overlap

The shared zone is real and it is usually where projects are won or lost: URL structure, load speed, headline copy, page hierarchy, forms, structured data. A designer who ignores those subjects delivers a beautiful invisible site; a marketer who inherits a badly built site spends the budget compensating for design flaws. That is precisely why WithNet keeps both skill sets in the same team.

Which one to call, and when

1. You have no site, or it is more than five years old: start with design. Advertising towards a page that does not convert is filling a leaking bucket.
2. Your site is decent but nobody finds it: that is a marketing problem, mainly search and content.
3. You have traffic but few enquiries: that is a conversion question, sitting between the two trades, journey audit, rewriting, testing.
4. You are launching a product on a short deadline: both, in parallel, with a media budget planned from day one.

The right reflex before signing: ask your contact what they accept to be judged on in six months. The answer will tell you, far more reliably than a brochure, which trade they actually belong to.

WordPress: the default choice, for good reasons

WordPress powers more than 40% of the web and the overwhelming majority of business sites in Tunisia. Its advantage is not technical, it is ecosystemic. You will always find an integrator, a plugin, a theme, an answer on a forum. The admin takes half a day to learn, which matters enormously when the person publishing is a sales assistant, not a developer.

• Best for: brochure sites, blogs, catalogues, shops up to a few thousand SKUs with WooCommerce.
• Weak spot: the ease of installing extensions turns against the site. Ten poorly chosen plugins are enough to double load time and multiply the attack surface.
• Maintenance cost: the lowest of the three, provided you deliberately cap the number of extensions.

Drupal: rigour, when the content structure is complex

Drupal thinks in content types, fields, taxonomies and granular permissions. Where WordPress would have you improvise, Drupal models cleanly: a member directory with fifteen roles, an institutional multi-site portal, a document-heavy intranet. It is also the strongest natively on security and multilingual handling : a genuine argument for a trilingual French, English and Arabic site.

• Best for: institutions, universities, NGOs, large portals, projects with a long life expectancy.
• Weak spot: the learning curve. Budget 40% to 70% more in development and training than a WordPress equivalent.
• Maintenance cost: high but predictable; major version upgrades are real projects in themselves.

Joomla: the historic middle ground, now losing momentum

Joomla sits between the two: more structured than WordPress, more approachable than Drupal, with multilingual support built in and no extension required. It is a decent tool, but its community shrinks every year. In practice, finding a provider in Tunisia willing to take over a Joomla site in 2025 is already hard, and will be harder in 2030. We no longer recommend it for a new build, only for maintaining a healthy existing one.

The CMS is not what determines your search rankings. Google indexes HTML, not software. A well-structured, fast, well-linked WordPress will always outrank a neglected Drupal, and vice versa.

How to decide in three questions

1. Who publishes day to day? If it is not a technical profile, WordPress spares you months of friction.
2. Does the content have a strong structure : relationships, roles, approval workflows? If so, Drupal earns back its premium.
3. What is your investment horizon? Beyond seven years, the local availability of skills weighs more than today’s feature list.

One last point that is often forgotten: whatever the CMS, insist that your content stays exportable. An accessible database, clean URLs and organised media guarantee that you can switch tools one day without starting from scratch. That portability, far more than the software name, is what protects your investment.

The fifteen-minute weekly routine

• Open the site on a phone, on mobile data, and check that the homepage renders in under three seconds.
• Send a test message through the contact form and confirm it lands in the right inbox : the most expensive and most silent failure there is.
• Clear pending comments and spam before it pollutes your indexation.
• Re-read the time-sensitive information: opening hours, prices, product availability, phone number.
• Glance at the 404 errors reported in Search Console and redirect the ones that still receive visits.

Technical maintenance: what actually breaks a website

In our experience in Tunis, four incidents account for nearly every emergency call we receive, and all four are avoidable with one hour of work per month.

• An expired SSL certificate. The browser throws a red warning screen and traffic drops to zero. Enable auto-renewal and set an alert fifteen days ahead.
• Un-updated extensions. They are the main entry point for code injection. Always update on a staging copy, never directly in production.
• A full disk. Automatic backups fill the server, the database stops writing, and orders quietly vanish.
• A forgotten domain renewal. An expired domain can be recovered, but email and website stay unreachable for days.
• Uncompressed images uploaded by an internal team: a single 6 MB photo in an article is enough to sink your mobile performance score.

A backup stored on the same server as the website is not a backup. Insist on an external weekly copy, at least four retained versions, and one tested restore per year.

Content: the only lever that really moves a website forward

Technically, a healthy site almost maintains itself. What makes it climb is regular publishing. Two genuinely useful articles a month, answering questions your customers ask on the phone, are enough to double the indexed surface of a brochure site within a year. Length is not the point: a page that precisely answers "how much does… cost" or "how do I choose…" keeps working for you for years.

Measuring without drowning in dashboards

1. Track four metrics only: organic sessions, top landing pages, submitted forms, calls placed from mobile.
2. Always compare year on year, never month on month : seasonality in Tunisia is brutal, especially in August.
3. Tie every article to the service page it is meant to feed, and watch the traffic of that page.
4. Hold a one-hour quarterly review instead of building a daily dashboard nobody will open.

Finally, put in writing who does what: who holds the credentials, who signs off on copy, who calls the host when things break. Most abandoned websites are not abandoned for lack of budget, but because nobody internally was formally responsible for them.

What a presentation package really contains

Behind the marketing label sits a precise technical scope: a page count, a design template, an admin system, hosting and a support window. Everything else (copywriting, photography, translation, campaigns) is almost always out of scope. So the first question is not how many pages, it is what will I have to produce myself.

The three tiers you will find everywhere

• The essential brochure site (5 to 8 pages). Home, services, about, contact, legal. It fits craftspeople, small practices and young companies whose priority is to exist on Google Maps and reassure a prospect who already knows them.
• The extended presentation (10 to 20 pages). One page per service, client references, a blog, often two or three languages. This is the format for SMEs that want organic search traffic feeding a long sales cycle.
• Bespoke development. Client portal, configurator, catalogue wired to your ERP, quotation engine. You do not buy it "just in case" : you buy it when an internal process is already burning expensive human hours.

Five questions to ask before you sign

1. Who writes the copy? If the answer is "you", add three weeks and a copywriting budget to your plan.
2. Can I publish a new page myself after delivery, without going back to the agency?
3. Does the package include technical SEO (tags, structured data, speed, sitemap) or only the design?
4. What happens after twelve months? Hosting renewal, migration rights, ownership of the code and the credentials.
5. How many design revision rounds are included? That single line is what derails eight projects out of ten.

A reliable test: ask to see two sites delivered more than a year ago under the package you are considering. If they are still fast and up to date, the provider goes the distance. If they are frozen or broken, the package is a loss leader.

The most common mistake: buying pages instead of outcomes

Comparing two quotes by page count is like comparing two cars by door count. What matters is the path a visitor travels between their problem and your enquiry form. A six-page site designed around that path converts better than a thirty-page site built to fill a menu. At WithNet we always start by listing the three actions a visitor must be able to complete; the page count follows from that list, never the other way round.

One last marker: your package must plan for what comes next. A frozen website loses its rankings within eighteen months. Decide at signature who publishes the next article, who updates the pricing, and how often. That decision is what turns a cost into an investment.

Editing creates pace, and pace is what retains

On the web, average retention collapses after thirty seconds. So the editor works by subtraction: cutting silences, breaths, warm-up sentences and transition shots that carry no information. A twelve-minute interview becomes a ninety-second sequence that is denser and more convincing than the original. That is not mutilation, it is writing.

Which formats to produce, and why

• One horizontal 16:9 cut of 60 to 120 seconds for the homepage and the about page.
• One vertical 9:16 cut of 20 to 40 seconds, derived from the same shoot, for social media and ads.
• Silent micro-loops of 5 to 8 seconds to illustrate a service page without distracting from the copy.
• A carefully designed still thumbnail: it decides the click long before the content of the video does.

Subtitles are no longer optional

More than 80% of videos start muted on mobile. Without subtitles your message simply does not exist. For a Tunisian audience, plan at minimum a French and an Arabic version; Arabic subtitling requires right alignment and a suitable typeface, otherwise legibility falls apart. Burnt-in subtitles work everywhere, but a separate caption file is preferable on your own site: it is indexable and translatable.

Never autoplay a video with sound. It is the number one cause of an immediately closed tab, and it directly penalises your user experience metrics.

Embedding video without destroying performance

1. Never load the video file on first render: show a light cover image and trigger the download only on click.
2. Host long videos on a dedicated platform; keep only short compressed loops on your own server.
3. Export in H.264 for compatibility and offer a lighter WebM version whenever possible.
4. Fix the player dimensions in CSS to avoid layout shift while loading.
5. Add a text transcript below the video: it serves accessibility, search engines and visitors in a hurry.

A well-edited video on a service page mechanically increases time on page and contact rate. A raw ten-minute file on autoplay does exactly the opposite. Between the two there is no hardware difference: there is editing.

1. Inform: the reference function

This is the foundation: who you are, what you do, where you are, at what price, within what lead time. A site that does this one job well already removes 60% of qualification calls. It requires clear pages, dated information, a complete contact page and compliant legal notices : the last point being mandatory and frequently neglected.

2. Capture: turning a visitor into a contact

• Forms for contact, quotes, callbacks : short, with one objective per form.
• Direct calls from mobile, with a tappable number present on every page.
• WhatsApp, the preferred channel in Tunisia, as a direct link rather than a heavy widget.
• Mailing list sign-up : the only audience you genuinely own.
• Online booking, which eliminates back-and-forth over availability.

3. Sell: taking payment without an intermediary

Catalogue, cart, online payment, order tracking, invoicing. In Tunisia, integrating a bank gateway (Clictopay) or a provider (Paymee, Flouci) requires a merchant file and an administrative delay that must be anticipated at the start of the project, not the day before launch. Cash on delivery remains dominant: your funnel must handle both cleanly.

4. Serve: the work after the sale

Customer area, case tracking, document downloads, knowledge base, complaint form. This function is the most underestimated, yet it is the one that most reduces your team’s workload: every frequent question documented on the site is a conversation that never happens on the phone.

A site doing all four functions badly is less useful than a site doing two of them perfectly. Choose your priorities based on what costs your company most today: prospecting, selling or support.

The invisible technical services that condition everything

1. Hosting and domain name: availability and speed depend directly on this choice.
2. SSL certificate: without it browsers block access and forms lose all credibility.
3. Backup and restore: the only real insurance against human error and hacking.
4. Technical SEO: URL structure, sitemap, structured data, multilingual tags.
5. Privacy-respecting analytics, without which no decision can be taken seriously.

In short: a website delivers four visible services and five invisible ones. The visible ones decide your commercial results; the invisible ones decide whether the visible ones still work in two years.

Visual hierarchy comes first

On a screen, the eye does not read, it scans. The designer’s job is to build a slope the gaze slides down naturally. That takes only three tools : size, contrast and space. A heading twice the size of everything else, a button in a colour used nowhere else, and emptiness around what matters are enough to guide any visitor.

Typography does ninety per cent of the work

• Two type families maximum. One for headings, one for body copy. Beyond that the site loses its identity and gains load weight.
• A scale, not random sizes. Define five or six coherent steps and never step outside them.
• Line length between 55 and 75 characters. The most broken and most profitable rule to respect: beyond it, the eye loses the start of the next line.
• Generous line height. 1.6 to 1.75 on body copy, and more still in Arabic, whose ascenders and ligatures need air.

Designing in Arabic: what genuinely changes

A trilingual site is not designed in French and then "translated". Arabic reads right to left: navigation, arrows, progress indicators and directional icons must flip. Elegant Latin typefaces have no automatic Arabic counterpart : you pick a compatible Arabic family, usually taller, and recalibrate the sizes. Arabic copy also runs 15% to 25% shorter than French, so layouts calibrated on French alone fall out of balance.

Always test your layout in greyscale. If the hierarchy holds without colour, it holds everywhere: on a badly calibrated screen, in direct sunlight, or for the 8% of men who struggle to separate red from green.

Colour is a functional tool, not decoration

A professional site runs on one neutral dominant colour, one accent reserved for actions, and two or three greys. The accent should appear only where you want a click. The moment it also decorates a heading or a border, it loses its pulling power. Check contrast systematically: 4.5:1 minimum for body text, 3:1 for large headings.

From visuals to conversion

1. Define the primary action of each page before opening a design tool.
2. Place that action above the fold, then repeat it at the bottom of the page.
3. Delete every element that helps neither understanding, nor deciding, nor acting.
4. Check the real rendering on a mid-range phone, not on a studio monitor.

Successful web design is rarely noticed. What the visitor notices is that they found what they came for without effort : and that is precisely the outcome of invisible but rigorous design work.

What Facebook does better than your website

• Discovery. The feed pushes your content to people who were not looking for you. No website does that.
• Instant social proof. Comments, reviews, shares: a prospect sees other real customers before they ever talk to you.
• Conversation. In Tunisia, Messenger is still the preferred contact channel, ahead of email and sometimes ahead of the phone.
• Entry cost. Posting is free, and local advertising remains cheap compared with other ad networks.

What Facebook will never replace

A Facebook page is rented premises: you own neither the walls, nor the customer list, nor the rules. An algorithm change can cut your reach fivefold overnight, a malicious report can suspend the page, and you cannot export your audience. A website belongs to you. It is indexed by Google, it appears on high-intent searches ("plumber Ariana", "website price Tunisia") and it keeps working on the days you do not post.

One simple test: if your page vanished tomorrow, how many customers could still find you? If the answer is "none", your business does not own its acquisition channel.

The real pattern: Facebook attracts, the website converts

The two tools do not compete, they chain together. The page creates discovery and maintains the relationship; the site provides detail, reassures and triggers the action. A hesitant buyer will always look for an address, legal details, prices and past work before committing. That information is unreadable in a chronological feed and perfectly at home on a website.

Five settings to connect the two

1. Put the site URL in the page action button, not only in the bio.
2. Point one post out of three to a specific page of the site, never to the homepage.
3. Install the conversion pixel so you can retarget site visitors in your campaigns.
4. Turn every question asked in the comments into a page or an article: it is content your audience has already validated.
5. Display Facebook reviews on the site and the site link on the page: the consistency reassures buyers and helps search engines alike.

In short: a Facebook presence matters, and can be decisive at launch, but it stays rented. The website is the only part of your online presence that nobody can take away from you.

What happens before reading

Pre-attentive processing (the stage that precedes conscious attention) automatically detects four things: contrast, size, orientation and movement. A designer who masters those four variables controls the path of the gaze without writing a single line. That is why one button in a vivid colour mechanically attracts the click, and why a page where everything is emphasised emphasises nothing.

Three well-documented effects

• The picture superiority effect. Information paired with a visual is retained markedly longer than purely textual information.
• Reduced cognitive load. A diagram often replaces three paragraphs of explanation: the reader grasps a relationship at a glance instead of mentally rebuilding a sequence.
• Credibility transfer. Careful visual execution is spontaneously read as a sign of professional seriousness, even when the underlying content is identical.

Why graphics sell, concretely

On a product page, replacing a generic photo with an image showing the object in use and at scale visibly increases the add-to-cart rate. On a service page, a three-step diagram does more for comprehension than six hundred words of copy. And on a pricing page, a legible table defuses price anxiety better than any sales argument.

Beware the opposite trap: a decorative visual carrying no information consumes attention and gives nothing back. The test is simple : if removing the image loses nothing, it should have been removed.

Using visuals without weighing the site down

1. Prefer SVG for icons, diagrams and logos: it stays sharp at every size and weighs a few kilobytes.
2. Reserve photography for real things (products, team, job sites) and ban interchangeable stock illustrations.
3. Always set image dimensions in CSS to avoid content shifting during load.
4. Describe every information-carrying visual in its alt text: what the eye sees must also exist for a screen reader and for Google.
5. Keep treatment consistent (same palette, same crop style, same grain) across every page.

Graphics do not have more impact because they are prettier, but because they arrive earlier. They occupy the space text cannot reach: the first three seconds, when the visitor decides whether to stay.

What a photograph is actually made of

• Light. Its direction, hardness and colour set the emotion. Raking light sculpts a material; frontal light flattens it.
• Framing. Deciding what enters the image, and above all what stays out. A good photograph is often one from which things were removed.
• Depth of field. It separates the subject from its background and directs attention without a single word.
• The moment. In reportage as in product photography, the gap between a good and an excellent frame is often a fraction of a second or a millimetre.

Interior and product photography: two demanding disciplines

Photographing an interior, a piece of furniture or a decorative object is not about taking a pretty picture, it is about rendering a material: the grain of wood, the weave of a fabric, the sheen of ceramic. It calls for broad, soft light, a perfectly stable camera and, almost always, several exposures blended to balance windows against indoor lighting. It is patient work, where observation matters far more than settings.

On a decoration or furniture website, photography is not packaging: it is the product itself. A customer will never buy a sofa they could not examine closely.

Why your photos decide your online credibility

A visitor judges a website in under three seconds, and that judgement lands on the imagery first. A recognisable stock photo sends an unambiguous signal: this company showed nothing real. A consistent series of genuine photographs (even imperfect ones) builds immediate trust. In the audits we run, replacing generic visuals with an on-site shoot is consistently among the highest effort-to-result interventions available.

Five rules for web-ready photographs

1. Shoot in indirect natural light whenever possible, near a window, without direct flash.
2. Keep a constant background and a single colour temperature across a series: consistency beats isolated perfection.
3. Leave empty space around the subject so square, 16:9 and vertical crops all come out of the same file.
4. Deliver files at 2000 pixels wide maximum, compressed to WebP or AVIF: beyond that you slow your site down without gaining perceived quality.
5. Name and describe every image. The file name and the alt text are read by Google and by screen readers alike.

Two hundred years after Niépce, photography remains the fastest way to make people understand what you sell. It deserves half a day of preparation just as a page of copy deserves a proofread.

Being found at the moment of decision

A customer searching "web agency Tunis" or "aluminium joinery manufacturer Sfax" is minutes away from a purchase. They are not scrolling a feed: they type a query and compare three results. Without a website you do not exist in that comparison, no matter how good your work is. That is the difference between an audience that finds you by chance and customers who are actively looking for you.

Establishing credibility that nothing else replaces

• An address of your own. A .tn or .com domain says the company exists and invests in its longevity.
• A professional email address tied to that domain, rather than a free account that raises doubts from the very first exchange.
• Legal details, a tax number, a physical address: exactly what every B2B buyer checks before signing.
• Documented past work, viewable without having to request a portfolio by email.

Working while you are not working

A website answers the recurring questions (hours, service area, pricing, lead times) around the clock. Almost all our clients report the same thing after three months: call volume does not drop, but the nature of the calls changes. Prospects arrive informed, the conversation starts further along, the sales cycle shortens. It is a time saving that is hard to quantify in advance and obvious in hindsight.

A website is not a digital brochure. It is the only commercial space whose rules, appearance and access conditions you set yourself : with no intermediary and no algorithm.

Not depending on a platform

Every company that built its business exclusively on a social page has one day faced a brutal reach drop, an account suspension or a rule change. The website is the only brick you genuinely own. It also lets you build a contact base that belongs to you and reuse it without going through a third party.

When a website is not the priority

1. If your order book is full for twelve months and you are not looking to grow, wait.
2. If your offer and pricing are not yet settled, clarify that first: a website will not compensate for fuzzy positioning.
3. If nobody internally can spend two hours a month on content, start with five solid pages rather than an ambitious project that will age badly.

Creating a website means deciding to be findable and verifiable. Everything else (design, technology, page count) follows from that decision, not the other way round.