What does AR and VR marketing actually cost?
Anyone who gives you a figure before asking how many products you want covered is guessing, and the guess will be wrong in whichever direction suits them. The total for this work is dominated by one variable that has nothing to do with software, and once you know which one, wildly different quotes stop being confusing and start being comparable. This page is about that structure rather than a headline number, because the number is mostly yours to determine.
Why do quotes for the same brief differ so much?
Because the brief usually does not specify the thing that drives the cost, so each agency assumes a different answer and prices their own assumption. One reads "AR try-on for our range" as a demonstration on three hero products; another reads it as every item in the catalogue. Both are quoting honestly and the totals can differ by fifty times.
The software is close to a fixed cost. Building a try-on experience that works, tracks reliably and runs in a browser is roughly the same engineering whether it serves three products or three thousand. What scales is content: each product needs an asset, and someone has to make it.
So the first thing to do with any proposal is separate the one-off build from the per-product cost. If a quote does not break those apart, it is not yet a quote.
What actually drives the number?
Five things, in descending order of impact. Most conversations focus on the last two, which is why budgets go wrong.
| Driver | Impact on total | Why |
|---|---|---|
| Number of products covered | Dominant | Each one needs a 3D asset or a garment-ready image set. This is a per-item cost and it does not fall much with volume. |
| Asset source | Very high | Existing CAD or 3D models cost almost nothing to adapt. Photography-only products need assets built from scratch, which is a different order of work. |
| Where it lives | High, and inverted | A product-page feature costs slightly more to integrate than a microsite and is amortised over years of traffic instead of weeks of a campaign. |
| Tracking type | Moderate | Surface and image tracking are well-solved. Body, face and garment tracking need more engineering and more testing on real devices. |
| Platform choice | Lower than expected | Browser delivery avoids app-store builds and review cycles, so it is usually the cheaper option as well as the one more people will use. |
What is the cheapest useful version?
One product, on the product page, with an asset you already have. If there is a single item that generates the most returns or the most pre-purchase questions, covering only that one answers the commercial question at the smallest possible cost: does this change behaviour for our customers.
This is worth more than a broader pilot, because the measurement is clean. One product page with the feature, compared against itself before and after, or against a comparable product without it, tells you whether to roll out. A microsite covering ten products tells you how many people clicked a link in a newsletter.
Any agency that will not scope a single-product build is optimising for their invoice rather than your decision.
Which costs appear after launch?
Adding products. This is the one that gets people, because it is invisible at signing and unavoidable later. Every new SKU needs an asset, and if the only route to creating one is back through the agency at their day rate, your running cost is permanently tied to their availability. Ask specifically whether your team can add products themselves, and what tooling that requires.
Device drift. Phone browsers change, camera APIs change, and an experience built two years ago will eventually break on a new OS release. This is real but modest, and it is the reason to prefer a supplier who still has the code rather than one who delivered a build and moved on.
Hosting and processing. For browser-delivered experiences on normal ecommerce traffic this is usually a minor line item. If a quote presents it as a significant recurring platform fee, ask what is being hosted that a CDN could not serve.
Is a fixed price or a day rate better?
Fixed price for the build, day rate or self-service for the catalogue. The build has a definable scope: this experience, on these pages, working on this list of devices. That is exactly the kind of work that should be quoted as a number, and an agency unwilling to fix it is telling you they cannot predict their own work.
The catalogue is different, because you will keep adding to it and you do not yet know how fast. Tying that to a fixed scope means renegotiating every season. Tying it to a day rate with no way for your own team to do it means paying that rate forever.
The arrangement that ages well is a fixed build plus a documented process your team can run, with the agency available for the items that need judgement.
What questions make two quotes comparable?
Five, and they take one email. How many products does this cover, and what is the cost of the next hundred? Which of our products already have usable 3D or CAD assets, and which need building? Does this run in a mobile browser with no install? What is the device list you will test against, and does it include a three-year-old mid-range Android? And who can add a product after launch, our team or yours?
Two proposals answered against those five become directly comparable, and the cheapest headline figure frequently stops being the cheapest total. The answers also tell you a great deal about who has done this before, because a supplier who has shipped this work has been asked all five and will answer without hesitating.
Common questions
- How much does AR marketing cost?
- The total is dominated by how many products you want covered, not by the software, because each product needs its own 3D asset while the experience itself is close to a fixed cost. Any figure quoted before that number is known is a guess. Separate every proposal into the one-off build and the per-product cost, and the quotes become comparable.
- Why do AR agency quotes vary so widely for the same brief?
- Because most briefs do not state how many products are in scope, so each agency prices its own assumption. One assumes a demonstration on three hero items, another assumes the full catalogue, and both are quoting honestly. Whether your products already have CAD or 3D assets is the second large variable, since building assets from photography alone is a different order of work.
- What is the cheapest way to test AR for our products?
- One product, placed on its own product page, using an asset you already have. Pick the item that generates the most returns or pre-purchase questions. The measurement is clean, because you can compare that page against itself before and after, which a multi-product campaign microsite cannot give you.
- What are the ongoing costs of an AR experience?
- Adding new products is the significant one, since each needs an asset. Ask whether your own team can create and upload them or whether every addition routes back through the agency at their day rate, because that single answer decides whether your running cost is bounded. Beyond that, expect modest hosting and periodic maintenance when mobile operating systems change.
- Is browser-based AR cheaper than building an app?
- Usually yes, on both build and reach. Browser delivery avoids app-store builds, review cycles and maintaining separate iOS and Android codebases, and it removes the install step that loses most of the audience before they ever see the experience.
