Most social teams launch a batch of videos, put money behind all of them, and wait to see what sticks. Kroger just showed there’s a smarter order of operations, and it’s one any dealership marketing crew can borrow, even without a grocery chain’s data budget.
- Kroger, Vidmob and MMA Global found that predictive creative scoring could forecast e-commerce conversion with 81% accuracy before any media ran.
- Ads that followed the model’s recommendations averaged four times better conversion rates, and ads built around people beat ads about products or shopping.
- Dealership social teams can use the same thinking to rank their videos first and spend second.
What Kroger Actually Tested
Kroger teamed up with creative data firm Vidmob and trade group MMA Global to answer a simple question. Can you tell which ads will sell before you pay to run them? The study looked at 1,934 video and image assets from Kroger campaigns on Meta and Google’s DV360. It judged each one on things like messaging, narrative structure, branding and whether real people interacted on screen.
The model learned from a full year of 2025 campaign data. Then it was checked against a separate dataset from the first quarter of 2026. That second step matters. It means the model had to call results it had never seen, which is a much tougher test than explaining what already happened.
The results were hard to ignore. Predictive scoring forecast e-commerce conversion with 81% accuracy. Creative that matched the model’s recommendations saw an average fourfold lift in conversion rates. And shifting media dollars toward higher scoring ads could potentially produce more than twice as many conversions from the same budget. The study also found that about half of Kroger’s spend on Meta and DV360 had gone to lower converting ads.
Moving Money Beats Moving Pixels
Vidmob founder Alex Collmer described two ways to improve creative. You can edit the ad itself. Or you can simply move your dollars behind the ads that are already better. That second option is the one most social teams skip.
Think about how a typical dealership month goes. Someone films a walkaround, a quick service tip, a customer delivery moment and a promo spot. All four get boosted with roughly equal money. A few weeks later, the team checks results and adjusts. Collmer’s point is that this old habit means spending blind first and fixing things later. If you can rank the videos up front, you stop paying to learn what you could have guessed.
He also pushed back on the habit of labeling creative as a “non-working” cost. Marketers often say creative drives 50% to 70% of results, so treating it as the wasteful part of the budget doesn’t add up.
People Sell Better Than Products
One finding should feel familiar to anyone who has posted car content. Kroger’s ads that focused on human experiences outperformed ads centered on products or the act of shopping. Grocery isn’t cars, so treat this as a strong hint for your showroom and not a hard rule. Still, it’s worth testing. A video of a family loading gear into a third row may beat a slow pan across the dashboard. A tech explaining a tire check face to face may beat a graphic listing service specials.
The broader lesson is to break your videos into parts you can score. Does a person appear early? Is the brand clear? Is there a simple story with a start and a finish? Is the message one idea or five? Once you tag those traits, your own past results start showing patterns.
Why This Gets More Urgent Every Month
Content volume is exploding. Collmer pointed to generative AI and creators as two forces pushing the number of ad assets up by 10 or even 100 times. More videos means more chances to waste money on weak ones. A signal before launch becomes more valuable as the pile grows.
He made another sharp point. If every competitor uses the same AI tools, the tools alone won’t give anyone an edge. What sets you apart is your own performance data. For a dealership, that’s the record of which of your videos turned into leads, calls and appointments. Nobody else has that.
Build Your Own Scorecard Before the Next Boost
You don’t need an enterprise platform to start. Pull your last few months of boosted videos and sort them by the result you care about most, whether that’s leads, form fills or service bookings. Tag each one for a handful of traits, like a person on screen, early branding, a clear story and a single message. Look for the traits that keep showing up at the top.
Then flip your process. Score new videos against that checklist before they go live, put most of your budget behind the top scorers and keep a smaller slice for testing new ideas. Kroger’s study suggests the payoff for picking first and spending second can be big. Your own numbers will tell you how big it is for your store.
