Webflow Pros and Cons: An Honest Developer's Review

I have shipped Webflow sites, and I have shipped Astro sites with Cloudflare Workers behind them. Both approaches matter at DignuzDesign because real estate clients have very different needs. Some want a polished, brand-led marketing site they can hand to a content manager. Others want a fast catalogue of properties with custom filters and a 3D viewer that has to load without choking the page. Webflow is excellent for some of these jobs and miserable for others.

The honest version of the pros and cons list is not "design freedom vs. learning curve." It is what you hit in week four, in month six, and on the upgrade invoice. That is what I want to write about here, because most Webflow articles read like a tour of the marketing site.

What Webflow actually gets right

This is where Webflow earns its price, and where I keep recommending it.

The output HTML is clean

The visual canvas produces semantic markup. When I inspect a finished Webflow page, the HTML is usually close to what I would hand-write: real heading hierarchy, real <section> elements, no soup of nested divs with inline styles. That matters for SEO and accessibility, and it means I can hand a Webflow site to another developer without apologising for the source. Most page builders cannot say that.

The CMS is a proper relational model

Webflow's collection system, with reference and multi-reference fields, is a real content model. For a real estate site with property listings, agents, neighbourhoods, and project phases, you can model each as a collection and connect them: one property belongs to one neighbourhood, has many photos, references one agent. The Editor interface is genuinely usable by a non-technical client, which is something I cannot say about most headless CMS dashboards. The handoff is real - the client logs in, edits their listings, publishes.

Hosting that I do not have to think about

Webflow runs on Fastly-backed global delivery with automatic HTTPS, immutable asset caching, and reasonable Core Web Vitals out of the box. For comparison, the Astro projects I deploy on Cloudflare also score well, but I built that pipeline myself. In Webflow, hosting is a checkbox. For a small studio handing a site to a non-technical client, that matters.

Where Webflow starts to push back

This is the part that does not make it into the marketing copy, and the part you should know before committing.

The CMS limits are real and they are what bites first

According to Webflow's official dynamic content limits, each collection is capped at 5,000 items, each project at a tier-dependent total (10,000 on Business, 20,000 on Premium), and each collection is limited to 5 reference fields. The single collection list display is capped at 100 items without pagination.

None of these sound restrictive in a demo. All of them have caught me on real builds. A property catalogue with multiple project phases, where each property references its neighbourhood, its agent, its floor plan type, its building, and its development phase, will burn through the five reference fields before you finish the schema. I have refactored Webflow CMS schemas mid-build to merge two references into a single field with conditional logic, just to stay under the cap. That is the kind of work that costs hours and produces nothing visible to the client.

Filter complexity hits a wall

Webflow's built-in filtering, plus the Finsweet CMS Filter library, handles "show me two-bedroom apartments under five hundred thousand euros in Lisbon" perfectly. It does not handle "show me anything within a radius of this map pin, sorted by walkability score, with the user's saved filter state restored from a cookie." For an MLS-style search that has to deal with thousands of listings, ranged filters, and persistent state, you end up writing significant custom JavaScript inside embed blocks. At which point you are paying Webflow's hosting fee for a single-page app you wrote yourself. Custom Webflow development can take you a long way, but at some point you are fighting the abstraction.

The third-party script tax

Webflow injects its own webflow.js runtime, plus animation interactions, plus whatever you add for analytics, forms, and integrations. The base bundle is not huge, but every interaction you add through the visual interactions panel ships JavaScript, even when the interaction is purely cosmetic. On a slow mobile connection in a regional market, this adds up. When I have built the same layout in Webflow and in Astro, the Astro version typically lands a meaningful margin ahead on INP, because it ships almost no JS by default.

Embedding gets quietly painful

Real estate sites need embedded content: virtual tours, 3D property viewers, calendar widgets, custom search components. Webflow lets you drop in custom code through embed elements, and that works for simple iframes. It gets ugly when the embed needs to talk to the rest of the page, share state, or load conditionally. We ship our own product, AmplyViewer - an interactive 3D property viewer embedded into real estate websites, so buyers can explore a property online before visiting. Embedding it into Webflow works, but anything beyond a basic instance, say wiring the viewer to a CMS-driven listings page so the right 3D model loads for the current property, involves enough glue code that you start to wonder whether the rest of the site should be in Webflow at all.

The pricing tier trap nobody warns you about

The site plan you launch on is almost never the plan you stay on.

According to Webflow's current pricing, the simplified plans introduced in May 2026 combine the older CMS and Business tiers into a single Premium plan, with 20,000 CMS items included. That looks fine until you do the math on a real estate brokerage with multiple agents, multiple developments, and a blog. A thousand properties, two hundred agents, a few hundred blog posts, plus all the image assets each one carries, and you are quickly into Premium or higher just to keep the catalogue live.

I cover the full breakdown in the real cost of building on Webflow, but the short version is this: budget the recurring hosting fee at the second-highest tier, not the cheapest, and budget for a workspace seat for every collaborator who needs to log into the Designer. Webflow's pricing math always favours the bigger plan once you have a real catalogue.

The handoff question

This is the question I ask before quoting any project: who owns this site in three years?

Webflow is a hosted platform. The CMS data is exportable as CSV, and the static pages are exportable as HTML and CSS, but the CMS-driven pages do not export with their bound data, and the design system does not transfer to another editor. You cannot pick up the exported folder and continue developing it as a normal project. If Webflow doubles its pricing or pivots away from your use case, the migration cost is real and it is not small.

For a real estate brokerage that wants a marketing site they can edit themselves for the next five years, Webflow is a strong answer. For a property developer building what is effectively a custom application - filterable inventory, sales-team logins, CRM integration, embedded 3D content - I now usually start the conversation with Astro plus Cloudflare D1, not Webflow. The build is more expensive up front and pays back over the lifetime of the project.

The other studio I run, Faraday3D, produces architectural visualisations that often need to live inside these sites for years. The handoff question is the same one we ask there: who maintains this asset, and can they keep it alive after we are done. Webflow's answer to that question is "as long as you keep paying us." Sometimes that is the right answer. Sometimes it is not.

When I still reach for Webflow, and when I switch

After all of that, I still recommend Webflow regularly. The job has to fit the tool.

Use Webflow when:

• The client will edit the site themselves and needs a Designer-grade authoring experience without a developer in the loop
• The content model is genuinely CMS-shaped: a manageable number of collections, fewer than ten thousand items across the project, and reference depth that fits inside the five-reference cap
• The site is mostly marketing pages plus a structured catalogue, not a custom application
• Time to launch matters more than long-term independence from the platform
• The team has a designer comfortable on the Designer canvas but no in-house developer

Switch to code, usually Astro on Cloudflare in my case, when the catalogue scales past Webflow's caps, when filtering or search needs to be smart, when you need to integrate tightly with another system, or when the client will fund proper maintenance. The Astro framework gives you the same clean HTML output Webflow produces, with no runtime overhead and full code ownership. For Jamstack-style property developer sites, I have written about why a code-first build often beats a CMS-first build in the Jamstack approach for property developer websites.

I lean on this same Cloudflare stack for my own products. AmplyDigest, the daily AI digest service that summarises newsletters and YouTube videos into one morning email, runs on Workers and D1. That kind of project would never live in Webflow. A clean real estate brochure site for a small agency would.

Frequently asked questions

Is Webflow worth it for a real estate website?

For a single-agency marketing site with a manageable property catalogue, yes. The CMS handles property listings well, the Editor lets a non-technical agent update content, and the hosting is solid. For a brokerage with thousands of listings, an MLS feed, or custom search, you will outgrow Webflow's CMS limits and end up patching with custom code. At that point a code-first stack is usually cheaper over the lifetime of the project.

What are Webflow's real CMS limits I should know about before starting?

From the official docs: 5,000 items per collection, plan-dependent total items per project, 5 reference fields per collection, and 100 items per collection list without pagination. The reference field cap is the one most teams hit first when modelling real estate data, because each property tends to need links to several other collections.

How does Webflow compare to WordPress for performance?

Webflow generally beats a typical WordPress build because it ships fewer plugin-driven scripts and runs on a managed CDN. The comparison gets closer when WordPress is run headlessly with a static frontend, but for the standard "WordPress with a page builder plugin" setup, Webflow wins on speed and on cleanliness of the output HTML.

Can I migrate off Webflow later?

You can export CMS data as CSV and static pages as HTML and CSS, but CMS-driven pages do not export with their bound data, and the design system does not transfer to another editor. Treat a Webflow build as a long-term commitment to the platform, not a portable codebase.

Should I use Webflow if I already know how to code?

If you code, you will sometimes feel constrained by Webflow's abstractions and pay for hosting you could provide yourself. For marketing-heavy sites where a client needs to edit content without a developer, Webflow still saves weeks of CMS scaffolding work. I use it for the right jobs, but I do not default to it.

When is Webflow the wrong choice?

When the site is really a custom application in disguise: a filterable inventory of more than ten thousand items, custom user dashboards, deep CRM integration, or interactive content that needs to share state across pages. Those projects are cheaper and more maintainable as code, not as a Webflow site with embed blocks holding them together.