How IKEA Scales AI to Power Competitive Pricing: From Data to Decisions
In a standout presentation at MRMW EU 2023, IKEA’s Oya Ünlü (Global Head of Data & Technology, Customer and Market Insights) and Ash Marriott (Product Owner, Price Competition) revealed how they’re leveraging AI and machine learning to stay ahead of competitors — by matching similar products at scale, delivering pricing intelligence in real time, and enabling action across teams.
Read the session’s summary or watch the full video presentation for free via the link below.
This article was generated using AI transcription and editorial analysis tools. While we’ve aimed for accuracy, minor omissions may occur.
Oya Ünlü and Ash Marriott share insights on a modern data architecture, smart algorithms, and agile collaboration that are transforming how IKEA makes pricing decisions across global markets.
Reframing a Manual Process with Machine Learning
Traditionally, understanding how IKEA products compared to competitor offerings involved a painstaking manual process. Local pricing or product teams had to sift through competitor websites, catalogs, and brochures to identify similar items. This process was slow, subjective, and difficult to replicate across countries, making it nearly impossible to scale consistently.
To address this, IKEA developed a proprietary solution that automates competitive pricing intelligence by combining data science and human expertise. The system begins by collecting product data through structured web scraping of competitor websites, capturing everything from names and descriptions to imagery and pricing. That data is then processed through machine learning models capable of identifying similar products based on both textual and visual characteristics. This combination of deep learning for image analysis and natural language processing for product metadata makes it possible to identify comparable items even when naming conventions or design cues differ across brands and regions.
The New Standard: Machine Learning at Scale
To address this, IKEA built a global, AI-powered system that automatically collects, processes, and analyzes competitor pricing data. The platform combines:
-
Automated data collection from competitor websites using web scraping
-
Machine learning algorithms that identify similar products by analyzing both text (e.g., product names, specs) and images (e.g., design features)
-
Similarity scoring models that provide confidence levels in the match
This allows IKEA teams to see at a glance how their prices compare to similar items — not just by name, but by actual design and market perception.
Human in the Loop: Trust Built on Transparency
AI is powerful, but trust matters. In the early phases, business teams manually validated the model’s outputs. As the system expanded, IKEA introduced a dedicated annotation team whose job is to:
-
Review and verify the accuracy of product matches
-
Improve training data with human context
-
Help the model continuously learn and improve
This “human-in-the-loop” approach ensures the models don’t operate as black boxes, and keeps confidence high within commercial teams who rely on the insights daily.
Seamless Integration Into Business Workflows
What makes IKEA’s solution stand out is not only the strength of its technology, but the way it’s been embedded directly into the company’s existing commercial workflows. The platform’s outputs are made accessible through internal dashboards that are tailored to the needs of pricing analysts, sales teams, and product managers.
These teams can view real-time comparisons between IKEA products and competitors in their market, including price differentials, product similarity scores, and historical changes. The insight goes beyond numbers: it offers context, helping teams understand where IKEA is over- or under-priced and why. This enables smarter, faster decisions grounded in data rather than assumptions.
Scaling Through Cross-Functional Collaboration
Behind the platform is a stable, long-term cross-functional team made up of data scientists, ML engineers, product owners, and insight specialists. Rather than outsourcing key components, IKEA chose to develop and maintain this system internally. This allows for tight integration with IKEA’s data architecture, as well as ongoing refinement based on feedback from commercial users.
The model is updated regularly based on new training data, and the annotation team continuously improves its quality by addressing nuanced or ambiguous cases. IKEA’s agile, iterative approach allows the platform to evolve in real time, adapting to changes in the market and to the company’s strategic priorities.
eating Organizational Buy-In for AI
A key success factor wasn’t the model — it was internal trust. To ensure widespread adoption, IKEA invested in training, documentation, and stakeholder engagement.
The team emphasized:
-
Clear explanations of how the model works
-
Transparent communication around strengths and limitations
-
Early involvement of commercial users in testing and validation
This helped teams across markets understand and believe in the system — not just use it.
Scalable Globally, Relevant Locally
While built for scale, the system is designed to reflect local realities. Each market can access insights that are tailored to their competitive landscape while feeding into a unified global architecture. It supports both strategic and tactical use cases, including:
-
Local pricing adjustments based on real-time competition
-
Global alignment on perceived value positioning
-
Product and range decisions based on market-level signals
This flexibility ensures that every team — from pricing to category to insight — can act confidently on shared intelligence.
Looking Ahead: Evolving with the Market
During the Q&A, Ünlü and Marriott confirmed that the platform is updated weekly, includes promotional pricing as a separate layer, and is already influencing decisions beyond pricing.
Potential future use cases include:
-
Monitoring design and trend shifts across competitors
-
Supporting IP and copyright protection
-
Feeding insights into product development cycles
As consumer expectations evolve, so too will IKEA’s use of AI to stay responsive, agile, and in sync with what value means in different markets.
📍Up Next: MRMW Europe Returns This Autumn!
Ready for more inspiration from brands like Porsche? MRMW EU is back on November 6 – 7, 2025 | Berlin, Germany — the must-attend event for insights professionals, marketers, and experience designers.
- Discover case studies, research trends, and industry breakthroughs
- Meet global brands and visionaries
- Network with the most influential voices in market research
Early bird pricing now available – save €400! Grab your chance to elevate your market research with the newest methodologies!