Konverta Lab · Buyer's guide
Shelf test and predictive attention tools compared (2026)
Which one fits your pack test? Twelve tools, from AI heatmaps to real shoppers on a virtual shelf. What each does, whether it handles a pack on a shelf, what's free, what it costs where that's public, and who it suits.
In short Predictive attention tools are fast and cheap, and good for screening early. Human-panel shelf tests are slower and cost more, and they're what you use to validate a final design. Our free Shelf Test sits at the screening end, built for packs on a shelf. For big decisions, put the finalists in front of real shoppers.
Last checked: 28 October 2026
01 Three kinds of tool
What kinds of pack test tools are there?
- Predict
Predictive attention (AI eye tracking)
Upload an image, get a heatmap of where eyes are likely to go in the first seconds. No people involved. Results in seconds. Dragonfly AI, Neurons, Attention Insight, 3M VAS, EyeQuant and Brainsight.
- Simulate
Synthetic shoppers
Predicted attention plus a simulation of shoppers moving their eyes around a shelf or an aisle. Our Shelf Test runs 1,000 of them, who differ in eye height, glance time and walking pace.
- Measure
Real shoppers on a virtual shelf
Real respondents see a shelf online or in a lab, and you record what they look at, click or buy. Ipsos Simstore, aytm, Toluna, RealEye and PRS IN VIVO.
02 The comparison
How do the tools compare?
Every fact below comes from the vendor's own site or from trade coverage, linked in the sources. Where a vendor doesn't publish something, we say "not public" rather than guess. Prices change; check before you buy.
| Tool | What it does | Pack on shelf? | Free option | Pricing (if public) | Best for |
|---|---|---|---|---|---|
| Synthetic shoppers | |||||
| Konverta Shelf TestKonverta Lab | Puts a pack on a supermarket bay and runs 1,000 synthetic shoppers on predicted attention: pack check, shelf visibility against competitors, shopper materials, aisle walk-by. | Yes, built for it | Yes. Full demo, plus one pack of your own each week, after a one-click email sign-in | Free. Your products on your real shelves is a custom build | Screening packs, shelf positions and shopper materials early, at no cost |
| Predictive attention | |||||
| Dragonfly AI | Creative testing platform built on a "biologically inspired" attention model, not trained on eye-tracking data. Pack testing, Amazon and retail marketing use cases. Partners with Mintel on a packaging score. | Yes. Shows pack analysis and shelf visibility scores | No. Book a demo | Not public. Plans tailored by sales | Brand and packaging teams wanting one enterprise platform across pack, shelf, e-commerce and ads |
| NeuronsNeurons AI (attention model called Predict in its knowledge base) | Predicts attention and scores creative, from a model trained on an eye-tracking database of over 20,000 participants. | Packaging is in its training data; no dedicated shelf feature stated | No. Demo and quote | Not public. Starter, Pro and Scale packages on quote | Ad and creative teams wanting a large eye-tracking-trained model with enterprise extras (SSO, brand kit) |
| Attention Insight | Heatmaps, focus maps, clarity score and AI recommendations, with Figma, Photoshop and InDesign plugins. | Any uploaded design; no shelf feature stated | 14-day free trial, no card | From €29 a month (Basic), €119 (Pro), €299 (Hero); Enterprise custom | Designers who want low-cost heatmaps inside the tools they already use |
| 3M VASVisual Attention Software | Predicts what viewers will see at first glance. 3M calls it "a visual spellcheck". | Any uploaded image; no shelf feature stated | Yes, two images free | Monthly and annual individual subscriptions; group licences for 20+ people | A quick first-glance check on any design, from a long-standing name |
| EyeQuant | Upload any visual and get a prediction of how people will see it. Web and ad focus; lists packaging among assets. Acquired under new leadership; its site says 2024, PitchBook lists October 2025. Buyer not public. | Packs, yes; shelf not stated | Free trial, no card | "Plans from $97/month"; tier prices on request | Website, landing page and performance-ad teams |
| Brainsight | Predicted attention for images, websites, HTML5 ads and video: heatmaps, gaze plots, attention and clarity scores. | Not stated; built for ads and UX | Free trial, no card | Team from €199 a month, Business from €399, Agency from €799; Enterprise custom | Agencies testing display, social, out-of-home and video ads |
| Real shoppers on a virtual shelf | |||||
| RealEye | Webcam eye tracking with real participants: your own, or its panel at $15 each. Virtual shelf testing module; shelf reports can add AI-predicted attention. | Yes, via the shelf testing module (sold as an add-on to any licence) | 7-day free trial | Basic $390 a month ($325 billed annually), Standard $990 ($825); shelf module price not public | Real eye tracking on a virtual shelf at a self-serve price |
| Ipsos Simstore | Virtual store research: 2D and 3D shelves, 360° store and e-commerce views, behavioural and attitudinal KPIs. Simstore Pack Express reports in 7 working days. | Yes | No | Not public | Later-stage pack validation with a global research agency |
| aytm Shelf Test | A simulated online shelf where respondents shop. Reports cart spend, time at shelf, time to cart and a click heatmap. | Yes | No | From $1,500 plus sample cost; needs a Max licence | DIY findability and choice tests with real respondents |
| Toluna Pack Test Express | DIY pack test on Toluna Start, using Toluna's panel and benchmark database. Timed exposure and a simple shelf context. Up to 12 packs per survey. | Yes, in a simple shelf context | Not stated | Not public | Fast pack screening with real respondents and norms |
| PRS IN VIVOBVA Group | Shopper and packaging research consultancy using eye tracking and behavioural science. AI Pack Screener 2 with Viz.AI predicts on-shelf visibility. | Yes | No | Not public | High-stakes pack redesigns that need eye tracking with real shoppers and an expert read |
Accuracy claims are left out of the table on purpose. Vendors measure them in different ways, so they don't compare. We unpack them in How accurate is predictive eye tracking?
03 Being fair
Where are the others stronger?
Our Shelf Test is free and built for the shelf. That doesn't make it the right tool for every job. Here's where others do more.
- Trained on, or measured with, real eyes. Neurons trains on an eye-tracking database of over 20,000 participants. RealEye and PRS IN VIVO measure real people. Our attention engine is built from published saliency methods and hasn't been validated in an eye-tracking study of our own.
- Real shoppers and real choices. Ipsos Simstore, aytm and Toluna put your pack in front of real respondents and record what they buy, with norms to compare against. Our synthetic shoppers predict who notices a pack, not who picks it up.
- Enterprise features. Neurons sells SSO, brand kits, workshops and a customer success manager. Dragonfly AI tailors plans to the business. Attention Insight and EyeQuant have Figma plugins.
- Formats beyond the shelf. Brainsight covers websites, HTML5 ads and video. EyeQuant covers websites, ads and emails. The Shelf Test is about packs, shelves, shopper materials and aisles.
- Words and brands. Our engine doesn't read words, recognise faces or know a brand. Real shoppers do.
Where the Shelf Test is useful: it's free, it puts the pack in shelf context against competitors, it adds shopper materials one at a time, and it walks shoppers down the aisle. With 1,000 synthetic shoppers, a result like "65% notice the pack" is precise to about ±3 points. That's the precision of the simulation, not its accuracy against real shoppers.
04 How to choose
Which tool fits your pack test?
- Early
Concept screening
Ten ideas, little budget, a week to decide. Use predictive attention or synthetic shoppers to catch problems and cut ten to two. Speed matters more than precision here. Our Shelf Test and the attention heatmap are free; several paid tools offer trials.
- Middle
Choosing between finalists
Two or three designs left. Run them all through the same tool and compare them against each other, not against an absolute score. If the gap is small, don't trust it. Move to real shoppers.
- Final
Validation
A redesign going to print across a range, or a launch with a forecast attached. Use real shoppers: a virtual shelf with a panel, or eye tracking in a lab. That's where Ipsos, aytm, Toluna, RealEye and PRS IN VIVO earn their fee.
When do you need real shoppers?
- When the decision is expensive to reverse: print runs, range-wide changes, new brand identity.
- When you need to know what people buy, not just what they notice.
- When brand familiarity matters. Loyal shoppers search for your pack; a heatmap doesn't know that.
- When the pack relies on words: a claim, a flavour name, a price message.
- When you need a number for a forecast or a board paper.
05 Questions
Pack test tools FAQ
Is there a free shelf test tool?
Yes. The Konverta Lab Shelf Test is free: sign in with a one-click email link and test one pack of your own each week. Several paid tools offer free trials instead: Attention Insight (14 days), RealEye (7 days), 3M VAS (two free images), plus EyeQuant and Brainsight trials. As of 28 October 2026 we found no free option for Dragonfly AI, Neurons or the human-panel shelf tests.
Which predictive attention tools handle a pack on a shelf?
Dragonfly AI shows packaging and shelf visibility use cases, RealEye sells a virtual shelf testing module, and the Konverta Shelf Test is built for packs on a shelf. Most other predictive tools analyse any image you upload, including a shelf photo, but are built mainly for ads, websites and design files.
Are predictive eye-tracking tools as accurate as real eye tracking?
Not on a real shopping task. Vendors claim 89% to 96% accuracy, but these are self-reported and usually measure first-glance, free-viewing attention, not a shopper searching for a brand. Use predictive tools to screen and compare designs, then validate finalists with real shoppers. The evidence is in How accurate is predictive eye tracking?
When do I need real shoppers instead of a predictive tool?
When the decision is expensive to get wrong: a final pack redesign, a range-wide change or a sales forecast. Real shoppers bring goals, brand familiarity and purchase behaviour that a predictive model does not see. Ipsos Simstore, aytm, Toluna, RealEye and PRS IN VIVO all use real respondents.
What is the difference between synthetic shoppers and a human-panel virtual shelf?
Synthetic shoppers are simulations: in the Konverta Shelf Test, 1,000 of them look around a shelf led by predicted attention, so a test runs in seconds and costs nothing. A human-panel virtual shelf shows the shelf to real respondents online or in a lab and records what they look at, click or buy. It takes days and costs money, but it measures real people.
06 Sources
Where these facts come from
All accessed 28 October 2026. Vendor pages describe the vendor's own product; we haven't tested the paid tools ourselves. Spotted something out of date? Email [email protected] and we'll fix it.
- Konverta Lab, Shelf Test and attention heatmap (model, references and limits).
- Dragonfly AI, home page, plans and our science; Mintel, Mintel and Dragonfly AI partner.
- Neurons, pricing, product page and validation and accuracy.
- Attention Insight, pricing and home page.
- 3M, Visual Attention Software and pricing and sign-up.
- EyeQuant, about, pricing and home page; PitchBook, EyeQuant profile.
- Brainsight, pricing.
- RealEye, pricing, virtual shelf testing and shelf report: predicted attention.
- Ipsos, Ipsos launches Simstore virtual solutions (2 April 2019), Simstore Pack and Simstore Pack Express.
- aytm, Shelf Test.
- Toluna, Pack Test Express.
- MrWeb Daily Research News, PRS IN VIVO launches AI Pack Screener 2 (8 November 2021). PRS IN VIVO's own site did not load on our check date.
Start with the free Shelf Test
Put your pack on a shelf and run 1,000 synthetic shoppers past it, in seconds. Need help choosing a research approach, or want your real products on your real shelves? Talk to us.