Development
What to Build Your Site On in 2026: Tilda, WordPress, Bitrix or Next.js
Builder, CMS or custom code is not a matter of taste but of scenario and planning horizon. We break down three-year cost of ownership, the ceiling of each platform, and what it costs to move if you picked wrong.
In short
The rule is simple: under ten pages with no integrations means a builder like Tilda, around 80,000-210,000 RUB over three years. A content site or blog means WordPress. A store tied to Russian accounting systems means Bitrix. A product, heavy traffic or non-standard UX means Next.js with a headless CMS. Getting it wrong is expensive: migrating between platforms costs 40-80% of a fresh build.
The short answer: platform by scenario
Arguing Tilda versus WordPress in the abstract is pointless, because they solve different jobs. The question worth asking is: what scenario does this site serve, how many people will work on it, and where will the project be in two years. Those three answers point to a platform almost every time, and after that you only need to check the budget.
The real divide is not builder versus custom code, it is showcase site versus working tool. A showcase describes the company and collects enquiries, gets edited once a quarter, and any platform handles it. A tool participates in the business process: it calculates prices, syncs stock, authenticates dealers, serves thousands of requests a minute. Builders run out of road here, quickly and painfully.
The second factor is who runs the site after launch. If that is a marketer with no technical background, the platform has to forgive mistakes: block editor, preview, version rollback. If it is an in-house team with a developer, flexibility wins, and a headless CMS with a custom front end delivers speed a block editor never will. The table below maps six common scenarios to concrete recommendations.
| Scenario | Recommendation | Why | Starting budget |
|---|---|---|---|
| One-person business or small studio site | Tilda or a similar builder | Fast, self-serve, no vendor on retainer | 20,000-60,000 RUB |
| Landing page for paid traffic | Builder on a small budget, custom code on a large one | Load speed and A/B freedom decide it | 50,000-150,000 RUB |
| Corporate site for a manufacturer | WordPress or Bitrix, Next.js for non-standard UX | Heavy content, editor roles, product catalogue | 150,000-500,000 RUB |
| Online store with accounting sync | Bitrix, or Next.js on its API at high traffic | Ready sync, fiscal receipts, local payment flows | 400,000-1,500,000 RUB |
| SaaS product or customer portal | Next.js plus a custom back end | Application logic does not fit a CMS | from 300,000 RUB |
| Media outlet, blog, knowledge base | WordPress or a headless CMS with Next.js | Editorial workflow matters more than layout freedom | 150,000-600,000 RUB |
Three-year cost of ownership, the part comparisons skip
Almost every platform comparison stops at launch cost, which is the least relevant number available. A site lives three to five years, and over that span subscriptions, licences, renewals, hosting and developer hours add up to a figure that often exceeds the original budget. Below is approximate arithmetic for a mid-sized corporate site over three years, based on 2026 market prices.
Tilda is the cheapest over time as long as the site stays inside its limits: subscription plus domain is tens of thousands of rubles across three years, with a one-off build from 50,000 RUB. WordPress looks free, but the real money goes to hosting, paid plugins and above all maintenance: core and plugin updates, cleanup after intrusions, speed work. Bitrix needs a licence plus renewal, roughly 20-25% of the licence cost each year, plus a developer who knows that specific system.
Next.js runs the opposite curve: high one-off entry, low annual cost. No licences, no plugins to patch against vulnerabilities, static hosting costs almost nothing and traffic scales without a rewrite. Over one year it is the most expensive option. Over three to five it usually lands close to Bitrix, and on an actively developed product it comes out cheaper, because each new feature does not mean fighting somebody else architecture.
| Platform | One-off at launch | Per year | Three-year total |
|---|---|---|---|
| Tilda | 50,000-150,000 RUB (build and design) | 10,000-20,000 RUB (subscription, domain) | 80,000-210,000 RUB |
| WordPress | 150,000-400,000 RUB (theme, customisation) | 50,000-130,000 RUB (hosting, plugins, maintenance) | 300,000-790,000 RUB |
| Bitrix | 250,000-700,000 RUB (licence and implementation) | 60,000-180,000 RUB (renewal, hosting, developer) | 430,000-1,240,000 RUB |
| Next.js with a headless CMS | 300,000-900,000 RUB (development) | 30,000-100,000 RUB (hosting, CMS, support) | 390,000-1,200,000 RUB |
Where each platform hits its ceiling
The Tilda ceiling arrives the moment logic does. A non-standard calculator, a portal with roles, catalogue filtering across ten parameters, two-way accounting integration: these are either impossible or held together with third-party widgets that break on platform updates. The second ceiling is speed. A page built from thirty blocks with background video cannot physically be fast, and on mobile traffic that is money leaving the building.
The WordPress ceiling is plugins. The core is fast and solid, but a typical corporate site carries 25 to 40 plugins, each adding its own CSS, its own JS and its own attack surface. That produces two chronic problems: speed that cannot be fixed without a cleanup, and security, because the overwhelming majority of compromises come through an outdated plugin rather than the core. WordPress demands maintenance discipline, and without it the site degrades quietly.
The Bitrix ceiling is the cost of change. It covers Russian specifics better than anything else: accounting exchange, fiscal receipts, document workflow, ready sales scenarios. But any deviation from standard behaviour is expensive, because you write inside somebody else architecture and the developer pool is narrower than PHP in general. The front end is its own problem: stock templates look their age, and modernising them often costs as much as a separate project.
The Next.js ceiling is people. It is the best option for speed, flexibility and control, but it needs a developer within reach: the content model has to be designed rather than picked from a list, and without a headless CMS an editor cannot change even a homepage block. Companies that choose Next.js with no support plan end up a year later with an excellent site nobody can update. The fix is known: a headless CMS on top and readable internal documentation.
SEO by platform: what genuinely moves rankings
Start with the myth: platforms do not rank. Search engines evaluate the final HTML, speed, structure and user behaviour, not the CMS logo. Platforms do differ in how much they help or obstruct that work. Tilda covers the basics, headings, meta tags, robots, sitemap, and stops there. Fine-grained structured data, canonical handling and clean URLs for complex structures run into editor limits.
WordPress has always been strong on SEO thanks to its ecosystem: flexible URLs, full control over structured data, mature sitemap and redirect plugins. Bitrix covers what is needed out of the box, including multi-level catalogues and indexable filter pages, but it needs configuration or it breeds duplicates. Next.js gives maximum control: server rendering, precise per-route metadata, excellent Core Web Vitals, all of which has to be programmed rather than assumed.
One fact matters specifically for the Russian market: Yandex has moved fully to mobile-first indexing, meaning it evaluates the mobile version of the page. If mobile hides part of the content, changes the markup or loads twice as slowly, that is your real ranking position, not the desktop one. Google has used Core Web Vitals as a ranking factor since 2021, while Yandex counts speed indirectly through behavioural signals: a slow page produces bounces, and bounces pull positions down.
- Tilda: basic SEO covered, fine control over structure and markup is limited.
- WordPress: strong ecosystem, but plugins hurt speed and need periodic cleanup.
- Bitrix: everything is there, but unconfigured filters and pagination breed duplicates.
- Next.js: the highest ceiling on speed and markup, all of it built by a developer.
Migration cost: the "we started on Tilda, now what" trap
The most common story: a company assembles a site on a builder in two weeks, grows, hits the limits and wants to "just move it". There is nothing to move. Builders do not hand over structured data and templates in a portable form, and an export gives you an HTML snapshot from which neither the content model nor the logic can be reconstructed. In practice migration means building again: the design can be reproduced, everything else gets paid for twice.
Rough market figures: moving from a builder to custom development costs 60-90% of a fresh build, WordPress to Next.js runs 40-70% because content is structured and transfers through an API, and putting a new front end on an existing Bitrix back end is 30-50%. One line is mandatory in every scenario: SEO migration, meaning a redirect map, preserved URLs and transferred structured data. Skip it and a 30-60% traffic loss for two or three months is close to guaranteed.
The practical takeaway follows. A builder is an excellent start when you treat it deliberately as temporary and set a horizon, something like "one year to gather demand data, then move to custom development". The bad scenario is choosing a builder as a permanent home for a project that will obviously outgrow it in six months. Then you pay twice, and between the payments you lose time fighting limitations.
Who owns the site: code, data and the right to leave
Ownership sounds like a legal question and is settled technically. On a builder you own the content and the domain but not the platform: while you pay the subscription the site works, and when you stop it goes dark. That is not malice, it is the service model, and it works perfectly for a showcase. The problem appears only when a meaningful share of revenue depends on that showcase and there is no plan B.
On WordPress, Bitrix and custom builds you hold the files and the database, which is the technical ability to change vendors. It is an ability, not a guarantee: if the code has no documentation, no version control and no coherent structure, the next team spends as much on understanding it as on rewriting it. So the contract should cover not only rights to the code but the obligation to hand over the repository, documentation and deployment instructions.
The practical minimum worth demanding on any project: repository access from day one rather than after the final invoice, editable design source files, a description of environment variables and external services, and deployment instructions a third-party specialist can follow. That package costs nothing when requested at the start and a great deal when remembered during a vendor breakup.
How to make the decision in one evening
Take a sheet of paper and answer five questions. How many distinct page types do you genuinely need? Who edits the site and how often? Is there data exchange with external systems, and in which direction? What traffic do you expect in a year? How much are you willing to pay annually rather than once? That last question kills the most illusions: people accept 700,000 RUB once and refuse 120,000 RUB a year, even though the second number is smaller.
If every answer is modest, under ten pages, edits once a month, no integrations, a few thousand visits, take a builder and stop paying for flexibility you will never use. If at least two items appear from the list of customer portal, accounting sync, filtered catalogue, non-standard pricing and heavy traffic, the builder is out and the real choice is between a CMS and custom development.
The final marker is horizon. A project with a one to two year life is almost always better on a ready platform, which will have earned its keep before it hits the ceiling. A project you plan to grow for three to five years and that earns money directly is better on a custom front end: the one-off cost is higher, but each additional feature gets cheaper rather than more expensive, which is the usual pattern when you build inside somebody else system.
Frequently asked questions
Can you build a real online store on Tilda?
You can, and for a catalogue of 100 to 300 items it works fine: cart, payments, simple filters and order export are all there. The limits show up beyond that. Complex multi-attribute filtering, per-customer pricing, two-way stock sync with accounting systems, loyalty programmes and non-standard delivery logic are either unavailable or delivered by third-party widgets you do not control. The rule of thumb: dozens of SKUs are fine on Tilda, hundreds of items with stock management need a CMS.
Is Bitrix outdated in 2026?
What is outdated is the stock templates and the "install it and never touch it" approach, not Bitrix itself. As a platform it remains the strongest option for Russian e-commerce: ready accounting exchange, compliant fiscal receipts, document workflow, mature order logic. Its weak points are the cost of non-standard changes and the look of boxed solutions. A modern working setup uses Bitrix as the back end and data layer with a Next.js front end on its API, which buys you local compliance and current-generation speed at once.
Is WordPress genuinely insecure?
It is not the WordPress core that is insecure, it is an unmaintained installation. The overwhelming majority of compromises come through outdated plugins and themes, weak admin passwords and missing two-factor authentication. A site with regular updates, a short list of plugins from active maintainers, a WAF and working backups runs for years without incident. The practical answer: budget for maintenance, 30,000 to 90,000 RUB a year for a mid-sized site, and treat every new plugin as a new employee with server access.
Is Next.js expensive only at the start or permanently?
Only at the start. The one-off cost is higher because you pay to build what a CMS ships with: content model, admin, roles. After that there are no licences, no paid plugins, no annual renewals, and static hosting costs far less than a full server running a PHP application. On an actively developed product the difference pays back in two to three years. If the site will not change for years, there is no saving to capture and a ready CMS makes more sense.
What should I pick when I cannot predict how the project will grow?
Split the site into two parts: the shell and the logic. Build the shell on whatever is cheapest and fastest right now, and keep the logic outside it, with payments, CRM, analytics and the database living in systems you can extract data from. Then changing platform means rebuilding the outer layer rather than losing everything you accumulated. It is the cheapest way to buy the right to be wrong: you lose a month of design work instead of two years of data.
Need a hand with this?
We do this work, not just write about it. Describe the task and we will scope it and send a staged estimate.
Related services
- Website development, end to endWe build sites that still hold up a year later: from a one-page CMS build to a store or a web application on React and Next.js. One team for design, code, integrations and launch.
- Landing page developmentOne page, one goal: the enquiry. We build the landing around a specific offer and the traffic you plan to send at it, whether that is search ads, paid social, organic or email.
- E-commerce developmentA store is not a shop window, it is an operational tool: catalogue, payments, delivery, stock and orders have to work as one. We build it so your team can run it without a developer.
Read next
- What a Website Actually Costs in 2026: The Estimate, Line by LineA practical breakdown of a web development estimate: what analytics, design, front end, back end and integrations actually cost, which expenses always show up after launch, and where cutting the budget is safe.
- A Slow Site Loses Money: Core Web Vitals and ConversionSite speed is not a technical metric, it is a line in the revenue report. Core Web Vitals explained in plain language, how to find the real cause of a slow page, and a fix list sorted by return per hour of work.
- How to Choose a Tech Stack: Criteria That Outlive the HypeA stack is a five-year commitment usually made in a single call. Here is the order the decisions should be taken in, the criteria that carry real weight, working stacks for six project archetypes and what a wrong choice costs.