Blog · November 3, 2026 · 7 min read

We simulated 5 live DTC ads. The same 4 mistakes kept showing up.

By Pavan, Founder, FirstReaction · Updated

In short

Across five live ads, the same gaps cost the most: no proof, no reason the price is worth it, a product never shown doing its job, and a headline that says nothing. Rebuilding each ad with only the brand's own claims raised predicted likelihood to act every time, by 8 to 24 points.

For the last few weeks we've been running a series called What's Wrong With This Ad? We take a live Facebook or Instagram ad from the Meta Ad Library, describe the people it's most likely aimed at, and simulate how they react on FirstReaction. Then we rebuild the ad using only the simulation's suggestions and only claims from the brand's own website, and simulate it again.

Five ads in, the interesting part isn't any single result. It's that the same few gaps kept showing up, in ads from very different brands, prices and countries.

The five ads

BrandThe adAudienceLikelihood to actRuns per version
Bella Vita"Buy any 4 perfumes for ₹1599", single imageYoung Indian perfume buyers, cold54 → 623
SnitchPlus-size jeans carousel, 5 cardsPast buyers and app users, hot61 → 736
Lola BlanketsRefinery29 feature, $325 blanket, single imageUS home and gift shoppers, cold39 → 603
Lola Blankets"Lola for the Holidays", 35% off codeSame audience44 → 683
Sahajan46-second creator video, $41 cleansing oilUS women 22–40 who wear makeup, cold59 → 673

Every rebuilt version scored higher. More useful than the scores, though, is why the originals lost people.

Mistake 1: no proof

Four of the five ads gave no reason to trust them. Bella Vita showed four bottles and a price, with no ratings. Lola's holiday ad had a mood, a dog and a discount code, but no reviews. Sahajan's video was a single creator's word. Lola's Refinery29 ad had press, but no sign of what thousands of buyers thought.

The fix in every case was proof the brand already had: Bella Vita's 4.8★ ratings, Lola's 4.9 stars from 39,729 reviews and its 35,000+ five-star reviews, Sahajan's 4.9★ from 257 reviews on a real customer's review card. None of it was new. It just wasn't in the ad.

Mistake 2: never saying why it's worth the price

In three ads, the buyer asking "is it worth it?" was among the least convinced. For Lola's $325 blanket, the Value-Driven Splurger scored 20 out of 100, the lowest of all. For Bella Vita, the Budget-Conscious Luxury Aspirant (45, second-lowest) wanted to know whether a budget perfume really lasts. Sahajan's simulation flagged "no direct value proposition" for a $41 cleanser. And Snitch's carousel showed prices to a hot audience of past buyers without an offer, so its deal-seeking buyers sat at 40.

A price on its own is not a reason. What worked: concrete details (double-sided, zero shed, machine washable), the scent family under each bottle, and, for an audience that already knows the brand, the brand's own app offer.

Mistake 3: the product never shown doing its job

Four ads showed the product without showing what it's like to have it. Lola's blanket appeared folded in a small frame, never in a home. Lola's holiday ad never said what Lola is. Bella Vita never said what the scents smell like. Sahajan's before-and-after was implied rather than shown.

The rebuilt versions put the product in use: a hand sinking into the blanket on a sofa, the scent family named under each bottle, the result on skin. If a stranger can't tell what you sell and why it's nice in a second, the rest of the ad doesn't get read.

Mistake 4: a headline that says nothing

Three headlines were just the brand name. "Snitch" scored 39 on FirstReaction's headline score. "Sahajan Skincare" scored 34. "Shop Lola Blankets" scored 40. Lola's discount-only headline, "Get 35% Off With Code: HOLIDAY35", scored 52. Headlines that said a benefit ("Life-Changing Softness", "Melts Makeup. Keeps Moisture.") or something specific ("Waist 40–50 · Flat 10% Off") scored 71 to 78.

There's a catch, though: fixing the headline alone barely moved anything. We wrote that up separately in brand-name headlines scored lowest, but fixing them alone didn't sell more.

Bonus: details that contradict each other

One ad had a problem no amount of creative skill would fix. Snitch's fourth card said ₹2099 on the image and ₹1469 in the headline, and its fifth card was an exact copy of the fourth. Shoppers read both, and a mismatched price reads as a trick. It's the cheapest fix of all: read every card against its copy before you launch.

What surprised us

  • The discount didn't carry the ad. Lola's holiday ad already had 35% off and scored 44. The offer helped once the ad said what Lola is and why it's worth wanting, not before.
  • The biggest gains came from the least convinced buyers, not the average. Snitch's deal seekers went from 40 to 82. More on that in design for your weakest buyer.
  • Fixes don't always add up. One round of changes left Sahajan's video flat; the second round worked. See fix, then test the fix.

A two-minute check before you launch

  1. Is there proof: a rating, a review count, a press mention?
  2. Does the ad say why it's worth the price, for someone who hesitates?
  3. Can a stranger tell what it is, and see it doing its job, in one second?
  4. Does the headline say something only your product can claim?
  5. Do every price and detail match across image, headline and copy?

If you want to run the same check on your own ad, the step-by-step version is in our guide to testing Facebook ads before launch. You can try it on FirstReaction free, with 25 credits.

Sources

  1. Meta Ad Library