When our marketing teams start new campaigns, we often begin by listening to what our customers are telling us. What do they love about the brand? What would get them to buy more, refer more and come back as a brand loyalist?
Recently, we’ve started looking more closely at another source of insight: our competitors’ customers. What are they complaining about? Where are competing brands falling short? And can we use those frustrations to identify opportunities for our clients to gain competitive share?
To accomplish this, our team at EGC Group had been developing an AI-powered Performance Media Engine (PME) to connect consumer research, competitive intelligence and advertising performance data.
The Performance Media Engine connects to our data sources through Model Context Protocol (MCP), which allows AI tools including Claude and Cursor to query live information from review platforms, forums, Reddit, etc.
For our client, positive reviews consistently highlighted two things: access to real clinicians who followed up and answered questions, and fast, discreet delivery. Customers described providers responding quickly, orders arriving within days, and a seamless checkout experience.
Competitor reviews revealed a different pattern. Common complaints centered on unclear checkout pricing, subscriptions renewing unexpectedly, and difficulty canceling or obtaining refunds. Phrases such as “charged again without notice” and “didn’t know what I’d actually pay” appeared across multiple competitors.
That contrast shaped the creative brief: emphasize what customers already valued about the client while directly addressing the billing concerns creating friction across the category.
The Before and After
Before the analysis, creative focused primarily on product features and convenience: how quickly customers could get a prescription online, and what promotional offers were available. Pricing and the checkout process were largely assumed to be understood, leaving two important purchase considerations underrepresented: trust and transparency.
After the analysis, messaging and imagery were rebuilt around three proof points: clinician support, fast and discreet shipping, and clear pricing without surprise charges. Headlines shifted from generic “get treated online” messaging to more specific claims about upfront and transparent pricing and real clinician follow-up.
Some concepts addressed common competitor pain points directly, without naming competitors, and contrasted those experiences with the client’s straightforward billing. Imagery also shifted toward the overall customer experience, making the delivery promise tangible rather than simply stating it.
With the help of the PME, EGC produces 50 to 100+ creative assets per month for this client, allowing the team to test the new messaging across multiple formats and concepts over a 15-day test period rather than relying on a single hero ad.
The results
Over the 15-day test, creative featuring the new messaging delivered 25% higher ROAS than the client’s existing creative, measured against the prior creative set as the comparison baseline. This was accompanied by a 10% improvement in blended account-level performance over the same 15-day period. (reflecting the broader impact using a multi-touch attribution model across paid social, search, and shopping.)
CTR and engagement also increased on ads highlighting the client’s value propositions and addressing common competitor pain points. This suggested that the messaging was earning attention and that the impact extended beyond impressions and clicks into conversion performance.
The Role of AI
Traditionally, this research would require manually reviewing customer feedback, categorizing themes in spreadsheets, and then investigating whether those themes correlated with actual ad performance. PME accelerated that process by clustering reviews into themes and cross-referencing them against creative already running in the account.
Our PME solution not only brought together public customer reviews but also competitor ad libraries, as well as the client’s own creative and performance data. It also incorporates anonymized, aggregated performance data from EGC clients in comparable categories to benchmark which messages tend to convert. Throughout the test, the PME allowed us to iterate hooks and CTAs of 100+ assets in real time.
Ultimately, our human team retained ownership of the strategy, creative direction, and final decisions. AI compressed the research and validation process so the team could move from customer insight to creative testing within a single sprint.
We went into this looking for ways to improve the creative performance. What we came back with was a better understanding of what mattered to the people seeing them. Having AI help with the research meant we could put those findings to work sooner.
