Reported execution: simplify the system before optimizing it
Microsoft says IKEA and Carat were managing campaigns and budgets across separate accounts. They brought product categories together in Performance Max, combined product photos with audience signals and rolled out additional campaigns across best-selling lines.
The human work remained upstream and around the system: deciding which products and audiences mattered, supplying the creative inputs and interpreting the resulting performance. AI handled more of the matching and delivery complexity once the structure was in place.
Our reading: automation is most useful after the brief is organized
The case is a reminder that an AI media layer cannot repair an incoherent product catalogue or an unclear conversion goal. Its advantage appears after the team has named the products, audiences, signals and measurement events that should guide delivery.
That is a quieter but more durable pattern than asking automation to invent strategy. The system can search a large decision space; the humans still define which outcomes are worth optimizing.
Production lesson: treat creative inputs as operating infrastructure
Performance systems need a complete, current asset set and a clean connection between the product feed, audience signals and conversion events. Creative freshness is not an afterthought when the platform is choosing among many combinations.
A practical review should look at which images and messages are being paired with which products, whether the landing experience keeps its promise and whether the automated winner is still on-brand. The dashboard is not a substitute for that inspection.
Evidence boundary: outcome figures need the missing denominator
Microsoft reports a 140% increase in return on ad spend, a 311% conversion-rate increase and a 215% increase in conversions. The public case page does not publish the comparison period, spend mix or independent validation.
This archive treats the case as evidence of a documented operating model—centralize, supply strong inputs and let the system optimize across signals—not as a guaranteed benchmark for another retailer.
Sources & limits
Microsoft Advertising case study dated 7 July 2026; ROAS, conversion-rate and conversion-volume figures are Microsoft/IKEA/Carat case claims, and the public page does not state the campaign flight dates.
- Microsoft Advertising — Modern Comfort Meets Modern Ad Strategies: IKEA’s AI Story — 7 July 2026.
- Microsoft Advertising — IKEA Customer Success Story PDF — source publication date not stated.
Send a correction with the passage and supporting source.