3D content used to mean one thing: a nicer-looking product page. Spin it, zoom it, maybe drop it into an augmented reality (AR) view. That’s still useful. It’s also no longer where the value is.
For CPG manufacturers and retailers, the same product model that runs on an ecommerce page can run on a virtual shelf test, a packaging review, a sales deck, and a shopper research study, without anyone rebuilding it from scratch each time. That works only if the model was built and stored with those other uses in mind, however, and most weren’t.
The questions retail teams are asking have changed. It used to be “Does this look good online?” Now it’s closer to “Will this package actually get noticed at eye level on a real shelf? Can we test a planogram before we ship products to 400 stores? Can the insights, ecommerce, and merchandising teams all work from the same model instead of three different ones that don’t quite match?”
That’s a different problem than the one most 3D programs were built to solve.
AR and VR Were the Entry Point, Not the Destination
AR and virtual reality (VR) are what got 3D content taken seriously in the first place. A shopper rotating a couch on their phone or walking a virtual aisle isn’t a gimmick; it cuts the uncertainty that drives returns, especially on products where size and fit are difficult to judge from a photo.
But the AR experience is the visible 5% of the work. Behind it sits the model itself, and behind every virtual store sits a library: fixtures, packaging, shelving, signage, and the environmental clutter that makes a simulation feel like a real store instead of a clean render. If that library is treated as a one-off deliverable for one campaign, most of the value evaporates once the campaign ends.
The 3D Assets You Need May Already Exist
Most CPG organizations already have product photography, CAD files, packaging assets, planogram models, and campaign renders scattered across a half-dozen systems and three agencies. Some of it is current. A lot of it isn’t. Teams routinely commission new 3D work because nobody can find the model that already exists, sitting in a folder two departments over.
That’s not a creative problem. It’s a governance problem, one that costs organizations money.
A 3D asset built for ecommerce can, in principle, be adapted for retail media creative, sell-in decks, packaging tests, shelf simulations, shopper research, training, and even a digital twin of a store environment. But that adaptation happens only if someone has tracked what the asset is, where it lives, who owns it, and when it was last updated. Without that, every team rebuilds.
Real-Time Rendering Turns 3D into a Decision Tool
The bigger shift isn’t improved visual fidelity. It’s the ability to rapidly render and modify output for testing. Move the package, change the lighting, drop the shelf position, and they can now test immediately.
That matters because retail doesn’t move on a quarterly cycle anymore. Packaging changes, retailer asks, new SKUs, and seasonal resets all land faster than the old render-and-review workflow can keep up with. A brand can now check how a new package reads at eye level versus a bottom shelf before committing to a physical mock-up or a field test that costs real money to run.
This is where 3D stops being a visualization tool and starts being a decision tool. It’s the difference between showing a stakeholder what something looks like and letting them find out whether it actually performs well before it’s sent to production. That distinction shows up in the numbers when teams run a virtual shelf test instead of guessing: One shelf test performed by InContext Solutions built around horizontal blocking surfaced a 14% lift for the client, which translated to $40 million in identified growth potential and an 800:1 return on the test itself. A separate packaging redesign study saved another CPG client $500,000 by confirming the existing design outperformed the proposed change, which is the less exciting outcome but the one that protects budget.
AI Won’t Save Bad Asset Libraries
AI is already useful here. It generates textures, cleans up models, converts formats, enables faster iteration on variants.
But what AI doesn’t fix is accuracy. Because the decisions downstream are real ones, a product model used for shelf testing or shopper research has to match the real product, dimensions, color, package hierarchy, claims placement, the works. If the underlying asset library is a mess of outdated, inconsistent files, AI doesn’t clean that up. It just produces more output, faster, from the same bad inputs.
The useful version of this isn’t AI replacing the people who build and manage 3D assets. It’s AI taking the repetitive part of that work off their plate so that they can spend more time on the parts that require human judgment.
3D No Longer Requires a Headset
A meaningful chunk of the shift here is the elimination of deployment friction. Interactive 3D models used to require a custom build or a headset. Now they run in a browser tab embedded in a product page, an internal tool, or a sales app.
For ecommerce, that means shoppers rotating a product without leaving the page. For sales teams, it means a customer conversation that isn’t a static deck. For merchandising, it means reviewing a shelf concept as something closer to the real thing instead of a sketch. None of this requires the shopper or the stakeholder to do anything unusual. It just works the way a web page works, which is exactly why the usage is spreading.
Digital Twins Make Retail Mistakes Cheaper to Find
A digital twin, in this context, is a virtual stand-in for a product, shelf, aisle, or store that lets a team test something before changing the real thing. This matters in CPG specifically because store resets are expensive, field tests take weeks, and historical sales data tells you only what happened under conditions that no longer exist once the shelf, the assortment, or the adjacency changes.
A virtual store lets a team run that test before the physical one. Not as a replacement for real-world data but as a way to narrow down which scenarios are even worth testing in the field. The credibility of that substitution depends entirely on how closely the virtual result tracks the real one. Virtual shopping research has been measured at a 96% correlation with actual in-store results, which is the number that makes “test it virtually first” a defensible strategy. Below that correlation threshold, a virtual test is just a more expensive guess.
The shelf is still where most CPG purchase decisions get made. A product isn’t competing in a vacuum. It’s competing against whatever else is on that shelf, at that price, in that store, for the three seconds of attention a shopper actually gives it. Bringing that context into the decision earlier, instead of finding out after the reset, is the actual value here, not the novelty of the simulation itself.
3D Strategy Has to Reach Beyond the Marketing Budget
3D gets framed as a conversion and engagement play, and it is one. But for CPG and retail organizations specifically, the larger value shows up in fewer physical prototypes, more consistency across channels, packaging testing before a production run is committed, faster retailer sell-in, and better decisions across teams.
The catch is that marketing is usually the budget owner, and marketing’s KPIs are about marketing’s use case. Insights, category management, sales, merchandising, and ecommerce may all benefit from the same asset library, but they rarely have equal say in how it gets built, governed, or prioritized. That mismatch, more than any technology gap, is why so many 3D programs stay siloed in one department instead of becoming the shared resource they could be.
Before building more models, the organization needs to answer a few unglamorous but important questions: which products need 3D assets first, which use cases actually matter, what level of detail each use case requires, who owns governance when five departments touch the same asset, how often models need to be refreshed, and who is responsible for noticing when they go stale.
Skip those questions and 3D becomes one more scattered-asset problem. Answer them and the same model pays for itself multiple times over.
Product Assets Become Infrastructure Only When They’re Built for Reuse
At InContext, we’ve spent more than 14 years building 3D retail environments, not just polished product renders. Today that work runs on a network of more than fifty 3D artists, with capacity to produce over 5,000 models a month, and a library of more than 400,000 models built to date.
Scale isn’t the interesting part, though. The interesting part is what the models feed into. Our 3D modeling work doesn’t sit in its own pipeline and hand off a finished render to whoever asks for it. It feeds directly into the same virtual environments we use for shelf testing, shopper research, packaging evaluation, and store simulation. Build the model once, and it shows up in a sell-in deck, a shelf test, and a packaging review without anyone remodeling it three times.
That distinction matters because a product-page render has only one job: look good. A model built for a virtual store has a more demanding role. It has to hold up when a simulated shopper walks the aisle, compares it against everything next to it on the shelf, and makes a choice. Most models never have to survive that kind of contact. Ours do, because that’s the only kind we build.
So the opportunity for CPG manufacturers and retailers isn’t really “get more 3D.” It’s building product assets that get reused across teams instead of rebuilt by each one. A model treated as a one-time creative cost dies with the campaign it was built for. A model treated as part of a managed library can support ecommerce, AR, packaging review, retailer sell-in, shopper research, and virtual store testing, all from the same file.
In our experience, most organizations are still running the first version, not the second. Getting to the second doesn’t start with building more 3D. It starts with an honest inventory: what already exists, what’s actually reusable, what needs to be rebuilt, and who owns the decision when five departments want to touch the same asset.
If you’re trying to figure out what building a reusable 3D asset library would look like for your team, let’s talk.
