1. The Setup
This report evaluates a single Meta in-feed image creative for buah, the freeze-dried-fruit DTC brand. The ad pairs an English-language lifestyle headline, "Sorry I can't. My hands are full," with German on-pack copy across four large-format cans featuring Erdbeere, Mangostan, and Pfirsich, set against a daylit domestic interior. It runs without a price, a call to action, a Bio-seal, or an influencer. Brand recognition rests entirely on the buah wordmark on the cans and the platform frame around the image. The test ran as a single-creative absolute-response simulation across 41 rounds on Meta, with Instagram as the primary placement and Facebook as secondary, targeting Germany as the primary market with German-speaking Austria and Switzerland in scope.
The central question was direct. Does this ad work, for whom, and why? More specifically, where does the lifestyle framing earn identity fit, where does the absence of product information stall momentum, and which objections point to fixable creative gaps versus deeper mismatches between the ad and how these audiences actually buy snacks.
The Ad Creatives
The creative is a single in-feed image. It leads with a relatable, English-language headline placed over a scene of large product cans in a domestic setting. The visual communicates the category: the cans display the product name and flavour variants in German, and the "#NICHTS ALS NATUR" (nothing but nature) and "100% FRUCHT 0% ZUSÄTZE" (100% fruits 0% added ingredients) claims are legible on-pack.
The question Imago set out to answer: could a simulation tell buah which audiences this creative is genuinely working for, which segments it leaves under-served, and what specific changes would close the gap between earned trust and completed purchase, before a single euro of real spend was committed.
How Imago Approached This
Imago's method is built around the idea that before you test an ad in market, you can test it against simulated audiences that behave like real people. Not simple A/B testing or focus groups. A simulation that generates the kind of reactions, hesitations, and language your actual customers would have when they scroll past your ad. Here is what that looked like in practice for this campaign.
1. Inputs
We fed the system the ad creative, the full copy, the brand context, and the campaign brief. This includes everything a real person would see: the image, the headline, the supporting copy, the product price range, and the platform context. The simulation needs enough signal to construct a credible impression of the ad as it would appear in a real feed.
2. Defining the Personas
Based on the brand context and campaign targeting, we defined four audience segments. Each one represents a meaningfully different type of consumer the ad was likely to reach in the German market.
- Family Snack Buyer: Parents shopping for children aged 4 to 7, running weekly lunchbox routines at dm or Rewe. Evaluates snacks on ingredient transparency, no added sugar, whether the product crumbles, and price per 100g compared to fresh fruit at Aldi.
- Urban Wellness Self-Consumer: City-based, active, label-reading. Reads Zutatenlisten before buying anything, shops at Bio-Markt and KoRo, and expects clean ingredients and sustainable packaging as a baseline rather than a premium.
- German-Speaking Snack Shopper: Based in Austria or German-speaking Switzerland, shops at Billa, Spar, or Hofer, and applies a practical, comparison-oriented mindset to snack purchases, checking unit prices and ingredient lists at the shelf.
- Plant-Based Active Adult: Vegan or plant-forward, fitness-active, Berlin or urban-based. Evaluates snacks on macros, ingredient sourcing, and packaging ethics, and holds new brands to a high evidentiary standard before switching.
3. Running the Simulation
200 AI agents, 50 per group, were exposed to the ad across 41 rounds and four exposure contexts: morning commute on Instagram, evening relaxation on Facebook, a parenting break on Facebook, and grocery planning on Instagram. Each agent could react the way a real person would: posting, commenting, liking, ignoring, clicking, or purchasing, with every reaction driven by how well the ad landed against their individual profile. Agents were not exposed to the ad in every round, mirroring real-world impression frequency, and the simulation tracked each agent's accumulating purchase intent over time, gating actual conversion on a prior click having occurred, matching how paid-ad attribution works in practice.
4. What the Simulation Captures
The output is two things working together. First, a set of quantitative metrics, CTR, engagement rate, purchase rate, sentiment scores, ROAS, and CAC, measured per segment and in aggregate. Second, and more importantly, the verbatim language agents used when they reacted: what stopped the scroll, what broke trust, and what question the ad consistently failed to answer. That combination shows not just which segment performed best, but exactly why the same creative that efficiently converts family shoppers actively alienates time-pressed young professionals, and what would need to change for that gap to close. That is what the following sections break down.
2. What the Simulation Found
Across all four groups and 41 simulation rounds, the buah "Sorry I can't. My hands are full" creative delivered a ROAS of 6.94 and a response rate of 18.1%, sitting well above the cold-audience norm of 1 to 3% and confirming the ad breaks through the scroll at a meaningful rate. Among agents who reacted, 74.8% of their actions were active rather than passive, posts, comments, and questions rather than idle scrolling, confirming that when the ad lands, it generates genuine deliberation rather than polite indifference.
The dominant emotional pairing across all groups was curiosity at 38.6% of classified felt responses alongside scepticism at 21.0%, and the single largest unspoken barrier was missing facts at 16.8% of all classified actions. Agents were drawn in and then left standing at the door.
The Four Groups at a Glance
| Segment | Purchase Rate | Positive Sentiment | CTR | ROAS | CAC |
|---|---|---|---|---|---|
| DE Family Snack Buyer | 16.0% | 38.6% | 4.8% | 11.11x | $2.25 |
| AT/CH German-Speaking Snack Shopper | 14.0% | 42.7% | 4.0% | 9.72x | $2.57 |
| DE Plant-Based Active Adult | 6.0% | 41.7% | 3.5% | 4.17x | $6.00 |
| DE Urban Wellness Self-Consumer | 4.0% | 48.6% | 5.0% | 2.78x | $9.00 |
The performance split across groups is where the simulation's most actionable findings live. The same creative that builds genuine conviction in one audience generates interest without resolution in another, and that gap carries real targeting implications.
What the Ad Got Right
The creative's core concept did genuine work. The lifestyle headline, the domestic setting, and the large-format cans together communicated a snack category quickly and clearly enough to generate above-benchmark response rates across all four groups. The concept of freeze-dried fruit as a convenient, portable, clean-label snack landed as personally relevant for audiences ranging from parents packing lunchboxes to urban cyclists managing gym bags, with "for me" reactions outnumbering "not for me" by a substantial margin across every segment.
The English-language headline in a German-market context is worth noting. Rather than generating alienation, it produced amusement in several groups, with agents reading the slogan as a relatable parenting or lifestyle moment. One AT/CH agent described it as feeling like "my life as someone who is constantly on the go." The mixed-language register created curiosity rather than confusion for the majority of agents, and that initial warmth is real capital the ad consistently earned.
"Freeze-dried fruit without additives? That's exactly the kind of snack I look for! Love having them on hand for that midday energy boost."
DE Urban Wellness Self-ConsumerWhat the Ad Struggled With
The single most consistent friction point across all four groups was identical: the creative withholds the information that DACH consumers need before switching to a new snack brand. No ingredient list. No price per 100g. No retailer signal. Agents across all segments raised the same questions in the same language, round after round, without resolution. Several noted explicitly that dozens of people were asking the same things across multiple threads and receiving no answer, and for the more criteria-driven groups, that silence gradually began to look like a deliberate choice rather than an oversight.
Beyond the information gap, price anxiety surfaced as a meaningful secondary barrier, particularly among Family Snack Buyers who benchmark everything against the unit price of a fresh apple at Aldi. A trust gap appeared specifically in evaluation-mode contexts, where the absence of any third-party credential left some agents unable to confirm the brand's clean-label claims. These are not fatal objections. They are solvable gaps that the current creative does not close.
How Context Shaped the Reaction
The simulation ran across four exposure contexts, and the differences in how the ad landed are worth noting. The same creative produced meaningfully different emotional registers depending on what agents were doing when it found them.
The morning commute on Instagram was the ad's strongest context by both response rate (21.5%) and CTR (5.8%). A commuter scanning fast reacts on instinct, and the buah visual and headline cut through quickly in that state. The primary barrier here was a trust gap rather than missing information specifically, which is a meaningful distinction: the commuter does not have time to interrogate claims, so the question is whether the brand feels credible enough to click on now. It mostly did.
The parenting break on Facebook delivered the highest positive sentiment of any context at 46.4% and the lowest negative at just 4.3%. The ad's humour and domestic setting landed softly with a tired parent taking five minutes away from the school run. This was also the only context where any share behaviour appeared, suggesting these agents were thinking about the product for others in their network. Despite the warmth, price anxiety appeared here in a way it did not in the commute context, pointing to a different kind of hesitation rooted in household budget logic rather than brand credibility.
The grocery planning context on Instagram produced zero ignored reactions and strong curiosity, but also the highest negative sentiment (10.4%) and the lowest CTR (3.3%) of any context. An agent actively planning this week's shop applies deliberate scrutiny to anything they encounter, and the creative could not clear that bar. Scepticism was elevated and a no-need signal appeared that was absent elsewhere: agents with a full shopping list already in mind are harder to interrupt with an unfamiliar category. The evening relaxation context on Facebook was the weakest. Indifference was the dominant felt emotion, the ignored rate was the highest of any context, and unclear product appeared as a barrier that did not surface anywhere else. Without an active mindset to decode the category, some agents were simply not sure what they were looking at.
What this tells buah is that the creative gap is not uniform across placements. In the commute context, a credibility signal may be enough to convert. In the grocery planning context, the audience is already in evaluation mode and only facts will move them. The remedy for each is different, and the placements should be treated as distinct briefs rather than interchangeable inventory.
"DE Family Snack Buyer"
This was the simulation's strongest-performing segment across every commercial metric. The highest purchase rate, the lowest CAC at $2.25, and a ROAS of 11.11x. But the number that matters most for understanding why this group converted where others didn't is the negative sentiment figure: just 3.4%. In a simulation where unanswered questions slowly turned curiosity into distrust for several other groups, this group largely held its warmth all the way through.
The dominant felt emotion across the run was curiosity at 42.2%, but what distinguishes this group is what sat alongside it: amusement at 10.7%, and very little distrust. Agents in this group laughed at the headline. The English-language slogan, "Sorry I can't. My hands are full," landed as a direct mirror of their daily reality, and that moment of recognition produced a warmth in the early reactions that few other groups matched.
That identity recognition is what drove the engagement rate of 81.7%, the second highest of any group. Agents who stopped scrolling were almost universally moved to do something active. The dominant impulse was to ask and comment, with agents debating the product in threads, swapping lunchbox tips, and discussing it alongside other snack brands they already trusted. The social behaviour had a communal quality to it that was specific to this group: agents were not interrogating the brand so much as talking to each other about whether it might work for their kids' lunchbox. That is a materially different dynamic from the scrutiny-driven comment behaviour seen in other segments.
The practical framing of the product was the emotional anchor here. Freeze-dried fruit as a solution to the mushy apple problem: fresh fruit that does not survive the school run intact. Agents returned to that framing across multiple rounds, unprompted, describing the concept as "handy" and imagining it fitting into a routine they already had. That specific use-case clarity, even without the product facts the ad withheld, gave agents enough contextual reassurance to sustain intent in a way that more criteria-driven groups could not replicate without the full information set.
The parenting break context on Facebook was where this group produced its most emotionally open reactions. Agents in that mindset were relaxed and low-scrutiny, and the ad's humour and domestic aesthetic landed at the right frequency. This was also the only context in the entire simulation where share behaviour appeared, with some agents wanting to pass the product on to others in their network rather than just engage personally. A tired parent recommending a snack to another parent is a different quality of signal than a curious shopper asking a question.
The frictions already covered in the overview section were present here too, but they operated as speed bumps rather than roadblocks. This group wanted the facts, but the lunchbox framing gave them enough to hold onto while they waited for answers that never came. Most did not convert despite that tolerance, which says something clear about how much more efficient this segment could be with the right creative additions.
"AT/CH German-Speaking Snack Shopper"
This group was the second-strongest performer in the simulation by purchase rate, and in many respects the most commercially interesting. A ROAS of 9.72x and a CAC of $2.57 sit close to the Family Snack Buyer's numbers, but the emotional texture of how this group arrived at those outcomes was meaningfully different. Where the Family Snack Buyer warmed quickly and stayed warm, the AT/CH Snack Shopper came in curious, stayed largely neutral, and converted through a more deliberate, comparison-oriented process.
The dominant felt emotion was curiosity at 37.8%, but what distinguishes this group from the Family Snack Buyer is the weight of amusement alongside it at 8.8%, the highest of any group. The slogan landed. Agents read the English headline and felt something, either amusement at the relatable chaos of hands-full daily life, or a lighter version of identity recognition that did not carry the same practical urgency as the parenting groups. One agent's early reaction captures it precisely:
"Haha, 'Sorry I can't. My hands are full', that's actually kinda funny. And the cans look nice."
That is a warm first impression but not yet a buying signal.
The 50.8% neutral sentiment is the defining feature of this group's emotional profile. It is the highest neutral share of any segment, and it reflects an audience that is genuinely undecided rather than passively cold. These agents were holding the product at arm's length while they tried to gather enough information to form a view. The engagement rate of 79.3% confirms that the minority who stopped were highly active, posting, commenting, and questioning in threads. The dominant social impulses were comment at 16.9% and ask at 13.9%, with zero save or send behaviour, a pattern consistent with publicly interrogating a new product rather than privately deciding to buy it.
One distinct finding is worth noting: in the very first rounds, one AT/CH agent flagged the mixed-language register directly, writing
"Wait, is this ad in English or German? The headline is English but the cans have German text. Confusing."
This is the only group where that friction appeared explicitly in the evidence. It did not suppress engagement or recur as a significant barrier across the run, but it is a signal that the mixed-language framing carries a small legibility cost with this audience that does not appear in the German groups.
The commute context on Instagram produced the strongest reactions from this group, with the slogan reading as immediately relatable to on-the-go shoppers navigating public transport. The grocery planning context, unsurprisingly for a unit-price-comparing Austrian and Swiss audience, produced the most critical reactions. Agents in that mindset applied shelf-comparison logic to the ad, benchmarking buah against Billa or Spar own-brand alternatives and arriving quickly at the same unanswered questions about price per 100g and retail availability.
"DE Plant-Based Active Adult"
This group is the most emotionally complex in the simulation. Strong identity fit, genuine enthusiasm in early rounds, the highest negative sentiment of any segment at 15.6%, and only 3 converted agents. The gap between how much this group wanted to like the product and how few of them bought it is the defining story here.
The dominant felt emotion was curiosity at 35.4%, but what made this group's early reactions distinctive was the quality of that curiosity. The product concept, freeze-dried fruit in a can, landed as a direct solution to a practical problem they already had: a clean snack that survives a cycling backpack or a gym bag without becoming pulp. One agent put it plainly:
"Freeze-dried fruit in my cycling backpack between classes? That would be a game-changer!"
Another described the packaging as "aesthetic," which for this group is not a trivial observation. Design is part of how they read whether a brand shares their values. The identity fit data reflects this: "for me" reactions outnumbered "not for me" by nearly three to one among classified responses.
The amusement signal was present but quieter here than in other groups at a moderate share. The English headline did not generate the same immediate laugh of recognition as it did with parenting audiences. This group's commute reactions were more considered from the start: the slogan was noted as "cute" but assessed rather than felt, which is consistent with a group who approaches everything, including ad copy, as something to be read and evaluated rather than simply experienced.
The trust gap at 3.1% of classified barriers is the figure that distinguishes this group from the others. It was the only group where this specific category appeared as a named friction source, and the appendix makes clear what it points to. Agents began to interpret the absence of facts as a deliberate signal. "It felt like greenwashing until proven otherwise" is the representative agent's phrase, and it appears independently in multiple threads.
"DE Urban Wellness Self-Consumer"
The dominant felt emotion was curiosity at 38.5%, but what sets this group apart from the others is the quality of that curiosity. Early rounds produced reactions that read less like consumer evaluation and more like genuine discovery:
"Okay 'buah' is such a cute name!! As a marketing manager in Berlin-Friedrichshain who bikes to work and does yoga in the mornings, I'm LOVING the concept of freeze-dried fruit that doesn't get squished in my cycling backpack."
That is someone who has already placed themselves inside the product's story. The identity fit data reflects this: "for me" reactions outnumbered "not for me" by nearly three to one.
Desire sat at a meaningful level in this group relative to the others, and it appeared alongside curiosity in a way that did not occur at the same intensity in the Family Snack Buyer or AT/CH groups. Agents were not just wondering whether the product might be useful but they were imagining owning it. One agent described the packaging as something that "would totally fit my kitchen shelf." Another said the concept was "exactly the kind of snack I look for." That is a different emotional register from practical consideration, and it explains the high CTR: this group was clicking because they wanted to find out more about something they were already drawn to, not because they had completed an evaluation and decided to proceed.
One friction unique to this group appeared in the very first active round of the run. An agent flagged the mixed-language register directly:
"Mixed language ads always make me skeptical, if you can't decide who you're talking to, what else are you cutting corners on?"
This reaction did not recur as a dominant pattern across the run, but it is the only group in the simulation where linguistic inconsistency was named as a trust signal. It surfaces an early vulnerability that the other groups did not express: for a group that reads everything with scrutiny, even small inconsistencies in creative execution become evidence about a brand's carefulness.
The engagement rate of 67.5% was the lowest of all four groups. The lower engagement rate reflects a higher rate of deliberate non-reaction: agents who saw the ad, formed a view, and chose not to engage publicly because they had already decided the ad had not given them enough to respond to. The agents who did engage were deeply so. The dominant impulses were comment at 12.5% and ask at 10.4%, with both save and send at zero. This group talked about the product in public rather than bookmarking it in private.
The evening relaxation context on Facebook produced the sharpest negative reactions from this group. Agents in a low-stimulus, reflective state applied the most rigorous scrutiny, comparing the brand explicitly against KoRo, denn's, and Bio-Markt alternatives they already trusted. In that context the ad's absence of a Bio-seal, retailer name, or ingredient confirmation left it unable to compete. The grocery planning context produced similar behaviour, with agents treating the ad as one item in a deliberate comparison exercise the creative was not equipped to win.
This is Imago
What a real-world campaign cannot tell you is any of the above. It can report that certain audiences didn't convert, but not that the DE Urban Wellness Self-Consumer arrived with the highest positive sentiment in the simulation and clicked through at a rate matching the best-performing group, yet stalled every time because the ingredient list, price per gram, and retailer were never confirmed; that the DE Plant-Based Active Adult entered each round actively wanting to find a brand worth supporting, felt genuine identity alignment with the product concept, and still left without buying and not from indifference but because the absence of a Bio-seal and a Zutatenliste read, round after round, as a reason not to trust; or that the DE Family Snack Buyer's warmth and durability across 41 rounds could have driven meaningfully higher conversion with a single price anchor and a retailer name on the creative. The simulation identified the segment worth leading with, named the one structural change, supplying the ingredient list, price per 100g, and a named retail channel in the ad unit itself, that would lift performance across all four groups, and gave buah clear grounds to understand exactly why, rather than having to spend months of real budget finding out, all before a single euro was committed.