We make a scrollytelling tool, so weigh what follows accordingly. It is still an honest guide, because the dishonest version is easy to spot and useless to you: there are projects where commissioning an agency is plainly the right call. Here is what an engagement involves, what it costs, how to pick a studio, and the one question that decides your route.
Start here, because the rest of the page only makes sense once this is settled. Five situations where a good studio earns its fee outright, and no tool closes the gap.
Bespoke art direction. If the piece must look like nothing else on the internet — illustration commissioned for this story, a motion language invented for this brand — you are buying taste, and taste is not a feature. Template-based systems, ours included, converge on a house style by design. That is the point of them and also their ceiling.
Original data journalism or reporting. Someone has to clean the dataset, interview the sources, check the claims and decide what the story is. That labour dwarfs the build. If your project starts with a spreadsheet nobody has interrogated, the interrogation is the project, and it is a human service.
WebGL, 3D and custom simulation. Scroll-linked 3D scenes, shader work, a globe you can spin. Specialist engineering: a generalist tool will not produce it, and a generalist developer takes three times as long as someone who does it weekly.
A one-off flagship with a real media budget. If you are spending six figures on distribution, the landing experience should not be the cheapest line item.
No internal capacity. The most common reason, and a perfectly good one. If your team is two marketers and a part-time designer, the fastest route to a finished thing may be paying people whose whole job this is. Tools do not create time.
If two or more of those describe your project, hire. Use the rest of this page to evaluate the quotes.
Most first-time buyers underestimate how much of an agency project is neither design nor code. A project runs through six stages, and only two produce anything you can look at.
Before anyone quotes, a studio needs to know how many sections, how much media, whether the content exists, whether there is a data pipeline, how many languages, and who signs off. Read the assumptions section of the proposal more carefully than the price — assumptions are where a fixed price quietly becomes variable.
Someone decides what the story is: the spine, the sequence, what the reader knows at each point. Good studios treat this as a distinct paid phase with its own deliverable — a narrative outline or wireframed scroll script — because approving the story before the visuals is far cheaper than approving it after. If a proposal jumps from kickoff straight to design comps, ask why.
Static comps of key moments, a motion study for the transitions, a design system for the repeating parts. Motion is hard to approve from stills, which is why better studios prototype one scroll sequence interactively before designing the rest. Ask for that.
Typically hand-written HTML, CSS and JavaScript, often with a scroll library doing the observation work. Scrollama, the MIT-licensed library maintained by Russell Samora, is a common choice in newsrooms and relies on IntersectionObserver rather than scroll events. Some of the same effects can now be done in pure CSS via scroll-driven animations, which run off a scroll progress or view progress timeline instead of a clock. Ask which approach a studio uses and what the fallback is, since support is not universal yet.
The stage that gets compressed when the schedule slips, and the one that decides whether the piece works on a four-year-old Android phone on a train. Ask what the test matrix is: which devices, which browsers, what network throttling, and whether they test with reduced motion enabled.
Contracts almost always name a number, commonly two or three. Anything beyond that is a change order at the hourly rate — reasonable, and also the mechanism by which a fixed-price project stops being fixed-price. Batch feedback and consolidate stakeholders before each round rather than during it.
Ask these four questions in the first call, not the last:
This is the honest strongest argument for using a tool, and it has nothing to do with launch day. It is about month seven.
A commissioned piece is usually a custom bundle: a hand-built page with a bespoke scroll timeline, tuned offsets, image sequences rendered at specific dimensions. It works beautifully. It also has no admin panel. So when legal asks you to soften a sentence, or a figure gets restated, or the person in the team photo leaves, the change does not go to whoever noticed it. It goes into an email to the studio, joins their queue, and comes back as a small invoice.
None of that is bad faith. It is the arithmetic of custom software, and it compounds:
Pages built this way tend to be published once and then frozen. Fine for journalism, quietly expensive for anything commercial — a report, a case study, a product story that has to track a product. If your content will change after launch, price the changes into the decision, not just the build.
The fix, if you go the agency route, is contractual: agree a support retainer or a per-change rate up front, and insist that whatever content is genuinely volatile lives somewhere your team can edit. Do not accept "we'll sort that out later." Later is when it costs the most.
We will not invent a price list. What we can do is point at a large public dataset of real agency engagements. Clutch's web design pricing guide, checked August 2026, reports an average reviewed project cost of $38,105.39, a typical project length of 7 months, and typical agency rates of $100–$149 per hour — while noting that most reviewed projects still come in under $10,000.
Those two numbers point in opposite directions, and that is the useful part. A typical project under $10,000 alongside a mean near $38,000 means a small number of large engagements are pulling the average up. Bespoke scrollytelling belongs in that tail, because the things that push a web project past average — original photography, commissioned illustration, custom motion, 3D, a data pipeline, localisation — are precisely what a scrollytelling brief asks for.
What drives your number, roughly in order of impact:
One practical tactic: ask for the quote broken into phases with a stop point after discovery. A studio confident in its process will price discovery separately and let you walk away with the narrative outline if the numbers do not work. One that insists on an all-or-nothing fixed price before knowing your content is guessing, and the guess will be padded.
Directories like Clutch let you filter studios by minimum project size before you spend a call finding out you are below it.
We deliberately do not publish a ranked list of agencies. We cannot verify other companies' client work, rates or capacity, and a list we cannot verify is worth nothing to you. The evaluation method is more durable anyway, because studios change.
Case-study pages are art-directed to flatter. Ask for live URLs of shipped work and open them yourself, on your phone, on a normal connection, ideally on a mid-range Android rather than the newest iPhone. If several portfolio links are dead, that tells you who maintains what after launch. Our gallery of examples and the roundup of famous scrollytelling examples are useful calibration before you review pitches.
Scroll experiences are heavy by nature, and the failure mode is jank under the reader's thumb. Interaction to Next Paint is a stable Core Web Vital, and the published threshold for "good" is 200 ms or less at the 75th percentile. Run a studio's live work through a field-data tool, then ask what the heaviest asset on the page is and how it is delivered. A studio that answers in specifics is one that thinks about it.
Two things to raise; the answers are diagnostic. First, prefers-reduced-motion, a CSS media feature available across browsers since January 2020, should already be wired into any motion-heavy build. Second, WCAG 2.2 Success Criterion 2.3.3, Animation from Interactions (Level AAA) requires that motion animation triggered by interaction can be disabled unless it is essential. Scrolling itself is fine — the guidance treats moving new content into the viewport as essential and user-controlled — but decorative parallax layered on top of it needs an off switch.
Ask past clients one question: what happened when you needed a change in month twelve? That answer predicts more of your experience than the whole portfolio does. Then confirm in writing what the editing surface is, who holds the hosting account, and how the assignment clause is worded.
If you are searching in German, the term is Scrollytelling Agentur, and plenty of searches arrive as Scrollytelling GmbH — the GmbH suffix simply reflects that the limited-liability company is the standard incorporation form in Germany, so buyers type the company designation into the query the way they would for any supplier.
The format has unusually deep roots in the region, and that shapes the market. After the wave of ambitious multimedia features in 2012 and 2013, the innovation team at German public broadcaster WDR built a reusable story engine so smaller newsrooms could produce the same thing without a bespoke build every time. That engine became Pageflow, open-sourced and introduced at re:publica in 2014, developed by Codevise Solutions and hosted in Germany — a detail the product still leads with, which tells you how often German-speaking buyers ask where the data sits. We keep a comparison of Pageflow and Scrollytelling if you are weighing that route.
The upshot: the same three options exist as everywhere else — commission an Agentur, license a platform, or use a browser-based tool — but the platform layer in the region is more mature than the English-language conversation suggests, and hosting location is a normal first-call question rather than an afterthought.
In practice the hire-vs-build framing is a false binary. Three routes exist, and they are good at different things:
| Commission an agency | License a platform | Browser-based tool | |
|---|---|---|---|
| Design ceiling | Unlimited — bespoke art direction, 3D, custom motion | High within the platform's system | Good, but converges on a house style |
| Time to a live page | Weeks to months, discovery first | Days to weeks after onboarding | Same day |
| Who edits it later | The agency, unless you contracted otherwise | Your team, in the platform | Your team, in the browser |
| Cost shape | Large one-off, plus change orders | Recurring licence, often annual | Low or usage-based |
| Best for | One flagship piece with a media budget behind it | Teams publishing regularly at enterprise scale | Frequent pieces, changing content, small teams |
The teams who get the most out of all three do some version of this: commission the flagship, standardise the rest.
The agency builds the one piece that has to be extraordinary — the campaign centrepiece, the anniversary story, the WebGL thing. Everything downstream, the pieces that change often, gets produced in-house with a tool: the interactive report, the launch page, the customer story, the recruiting page. The flagship sets the visual language; the tool carries the volume.
A second version is to buy art direction rather than production: commission a studio for narrative structure and design direction only — a scroll script, a type and colour system, a motion reference — and have your own team implement it. You pay for the expensive judgement and skip the expensive hours. It only works if your implementation surface can honour the direction, so check that first.
Either way, build a rough version yourself before you go shopping. Our guide to scrollytelling design patterns covers the structures nearly every piece is assembled from, scroll-driven storytelling covers the technical side, and the tools roundup compares platforms, including the ones that are not ours. If the rough version gets you most of the way there, the brief you write afterwards will be sharper and cheaper.
Usually a live URL and the files behind it: a designed, built and QA'd page, plus whatever source files and editing surface were in scope. What varies enormously is everything around that — who hosts it, whether a non-developer can change a headline, and whether the contract assigns you the copyright. Those questions decide what the piece costs you in year two.
Nobody can quote a real number without a scope. For a benchmark: Clutch's web design pricing guide, checked August 2026, reports an average reviewed project cost of $38,105.39, a typical project length of 7 months, and typical agency rates of $100–$149 per hour, while noting most reviewed projects come in under $10,000. Bespoke scrollytelling sits in the upper tail of that distribution, because original art direction, custom motion and heavy media are what push a web project past the average.
It is the German-language phrase for the same service — "Agentur" means agency, and German-speaking buyers search for "Scrollytelling Agentur" or "Scrollytelling GmbH", GmbH being the standard German limited-company form. The format has deep roots in the region: Pageflow, an open-source multimedia storytelling platform hosted in Germany, came out of broadcaster WDR's innovation team and was introduced at re:publica in 2014.
Whatever the contract says, and in the United States the default is less generous than most buyers assume. Under 17 U.S.C. § 101, work by an independent contractor only counts as a "work made for hire" if it falls into one of the enumerated categories and both parties sign a written agreement saying so. That is why competent agency contracts contain an explicit assignment clause. Ask for one, and ask separately about font, stock and library licences. General information, not legal advice.
Only if you specify it up front and test it before signing. A bespoke build is frequently a hand-written bundle of HTML, CSS and JavaScript with no editing interface at all, so every future change is a ticket back to the agency. If in-house editing matters, ask for a live demo of a non-developer changing a headline, swapping an image and publishing — during the pitch, on the real system.
When you need more than one of these a year, when the content changes after launch, when the deadline is shorter than a discovery phase, or when nobody on your team can maintain custom JavaScript. A tool caps how bespoke the result can be and in exchange removes the dependency: the person who wants the change can make it. For a single flagship piece with original reporting and a real media budget, an agency is still the better answer.
Build a working version of your story in the browser first. Either you ship it, or you hand a far sharper brief to the studio you hire.
Create your story → Browse templates →