Relevance & Research Question
Raiffeisenbank sought to optimize its loyalty programs to increase customer engagement and strengthen loyalty. The goal was to identify customer preferences and features that could drive higher program adoption and satisfaction.

Methods & Data

To address this, Lakmoos AI designed a targeted survey combining synthetic data and customer segmentation analysis:

1.Targeted Survey Design: The survey explored customer opinions on current loyalty features, preferences for new incentives, and variations in preferences across demographic groups.

2.Synthetic Panel Data Simulation: Simulated customer profiles provided predictive insights into how different demographics might respond to features like cashback, reward points, and exclusive benefits.

3.Real-Time Data Validation: Survey findings were validated against real customer data to ensure insights were accurate and actionable.

4.Quantitative and Qualitative Integration: Alongside quantitative questions, open-ended prompts offered qualitative insights into customer priorities and suggestions for program enhancements.

Results

Key insights revealed:

• High-Value Features: Cashback and flexible point redemption options were highly appealing across demographics, reflecting a preference for tangible rewards.
• Segmentation Preferences: Younger customers preferred gamified elements (e.g., badges, status levels), while older customers favored straightforward, financial-based rewards.
• Personalization Desire: Many customers indicated greater engagement when programs offered personalized rewards based on spending habits and lifestyle.

Added Value

The findings led to impactful changes for Raiffeisenbank:

1.Program Redesign: Enhanced loyalty programs now include more cashback options, personalized recommendations, and gamified features for younger users, boosting enrollment and engagement.

2.AI Integration: The successful pilot led to AI-driven research being integrated into Raiffeisenbank’s standard processes, enabling faster and more precise customer insights.

3. Synthetic Panels in Design Sprints: The adoption of synthetic panels in research democratized insights, making design sprints more inclusive and efficient.

These measures strengthened Raiffeisenbank’s ability to create customer-centric, innovative loyalty programs, enhancing their competitive position in the market.

Relevance & Research Question

Yahoo News sought to validate and size user personas based on preexisting qualitative findings to optimize product and marketing strategies. With a strong user base and brand identity, Yahoo News aimed to drive revenue growth. However, the team lacked a unified understanding of core users, which slowed innovation and delivery. They initially developed five personas through qualitative interviews but needed quantitative validation to determine their accuracy and distribution. Factworks was engaged to refine and enrich these personas for strategic prioritization.

Methods & Data

To quantify personas, Factworks translated qualitative insights into a structured survey administered to a representative Yahoo user sample. Instead of traditional segmentation techniques, the k-nearest neighbors algorithm was used to classify users based on similarity to preidentified personas. Respondents too dissimilar to any persona were excluded.

This approach allowed Yahoo to size each persona group and prioritize them strategically. Statistical testing helped identify key distinguishing traits, creating more distinct personas. Additionally, a Typing Tool was developed, enabling persona classification via a short questionnaire in Excel for future use in individual and batch scoring.
Results

• Yahoo teams gained a clear, shared understanding of their audiences through the study’s active socialization.

• Personas informed product, design, and editorial strategies across six News squads.

• They played a crucial role in cross-functional brainstorming sessions and roadmap planning.

• Product, design, and editorial roadmaps were prioritized based on persona insights.

• The study extended beyond News, influencing Yahoo’s Search and Games teams.

Added Value

Yahoo applied these insights to redesign and modernize its Homepage, tailoring updates to the three target personas. Results from iterative experiments included:

• 90% decrease in bounce rate

• 37% increase in classic page views

• 51% increase in Daily Active Users (DAUs)

Following the launch on June 13, 2024:

• Desktop experience saw a 30% increase across sessions, interactions, and page views.

• Mobile experience saw a 40% rise in user sessions and an 8% increase in total page views.

Relevance & Research Question
Our innovation, Intelligent Documents, redefines the delivery of research insights by transforming static reports into dynamic, interactive resources that allow end-users to engage directly with data. These documents are not just comprehensive reports—they are intelligent interfaces that respond to user questions, enabling follow-up inquiries and deeper exploration of insights. Researchers can navigate findings, request additional analyses, and validate each step of the analysis directly from familiar formats such as Microsoft Word, PowerPoint, or PDF.
Unlike traditional reports, which present insights in a fixed format, Intelligent Documents encourage real-time engagement. Every insight is backed by detailed validated data, which researchers can explore further by clicking embedded links that lead to the analytical foundation of each finding. For example, a marketing director or policy expert might ask, "What other factors are driving this trend?", "How does this vary across demographics?" or “How specifically were these results derived?” and receive immediate, data-driven answers within the same document. This interactivity is facilitated by utilizing Large Language Models (LLMs) to apply Retrieval Augmented Generating (RAG) to a selection of findings that are pre-generated for each analytical result, essentially as “speaker notes” for an LLM that guarantee verifiable answers free of hallucinations.
This interactivity elevates research from a static report to a dynamic platform for investigation, empowering end-users to make informed decisions on the spot. At the same time it combines “old” and “new” by embedding AI features into current document-centric workflows, thereby making the new capabilities of AI more accessibly and thus more relevant for day-to-day use.