Company
#
Chatbot
#
Live Chat
How Mars Petcare Built a Meaning Layer for Customer Conversations
About the Company
Mars Petcare is one of the world’s largest pet nutrition and care businesses, operating as part of Mars, Incorporated, a global consumer packaged goods company with over $50B in annual revenue.
Mars Petcare manages dozens of well-known brands including Pedigree, Whiskas, Royal Canin, IAMS, Cesar, Sheba, and Banfield, serving millions of pet owners globally across food, veterinary, and direct-to-consumer channels.
As a modern CPG and retail-driven organization, Mars Petcare operates at a massive scale across regions, languages, and product lines. Digital customer engagement plays a critical role in how the company supports customers, gathers feedback, and improves product and experience quality across markets.
Before Dimension Labs
Across its business lines, Mars Petcare generates large volumes of customer language through Ada chatbot conversations, customer support calls, and customer surveys.
These channels work well for engaging and supporting customers, but the data they produce exists as raw, unstructured text. Conversations could be manually reviewed or sampled, but they were difficult to analyze at scale.
Before Dimension Labs:
Chatbot transcripts and survey responses were unstructured and siloed
Analysis relied on manual review, keyword lists, and subjective sentiment scoring
Insights varied by team and market with no consistent logic
Two years of internal BI development failed to deliver scalable language analytics
Mars had access to the raw data, but not infrastructure required to extract meaning and insights.
With Dimension Labs
“Having customer language structured and reusable is foundational for how we operate.” — Digital Experience & Analytics Leader, MARS Petcare
Mars Petcare implemented Dimension Labs as the Meaning Layer for customer language, transforming raw conversations from Ada chatbots, support interactions, and surveys into structured, analytics-ready data.
Every conversation is analyzed at the record level and enriched with governed dimensions, creating a shared data schema for analysis across brands, regions, and time.
“What used to take days of manual work per region is now available instantly.” — Customer Analytics Leader, MARS Petcare
With Dimension Labs, Mars Petcare:
Structured every customer conversation into consistent signals
Unified chatbot, survey, and support analysis under one logic
Enabled cross-brand and cross-market comparability
Embedded conversation insights into weekly operating workflows
Removed manual review and ad hoc tagging
This foundation delivered measurable outcomes:
50% fewer chatbot escalations by fixing the flows that drove handoff
18% higher engagement through targeted optimization
~3 FTEs eliminated by automating conversation analysis
Faster insight cycles, shifting from monthly to weekly reporting
Customer language became a governed, operational data asset—powering continuous improvement across support performance, product feedback, and customer experience.
Use Case 01
Chatbot Performance and Escalation Intelligence
“We now have consistent visibility across chatbot, chat, and survey data in a way we never had before.” — CX & Insights Executive, MARS Petcare
Mars uses Dimension Labs to analyze customer conversations generated by the Ada chatbot across multiple brands, regions, and bots.
Instead of relying on high-level chatbot metrics like deflection rates or conversation volume, Mars structures every conversation into consistent, analytics-ready signals that explain what actually happened during the interaction.
This includes:
Why the customer reached out
What topic or category the conversation falls into
Whether an issue occurred and how it was handled
Whether the conversation escalated to a human
Whether the interaction was successfully resolved
With this structure in place, Mars can:
Compare chatbot performance across brands, regions, and markets using the same definitions
See which topics most often lead to escalation or drop-off
Identify gaps in chatbot coverage before they show up in CSAT or NPS
Prioritize improvements based on real customer behavior, not sampled transcripts
This shift allowed Mars to move from monitoring chatbot usage to actively improving chatbot effectiveness, contributing to a 50% reduction in escalations and a 18% increase in engagement within six months.
Use Case 02
Loyalty, Retention, and Marketing Enablement
“Dimension Labs has enabled us to move beyond anecdotal feedback and into measurable, operational insight.” — CX & Insights Executive, MARS Petcare
Beyond support and feedback, Mars uses Dimension Labs to extract early indicators related to loyalty, retention, and upsell opportunities directly from customer language.
Customer conversations are analyzed to surface signals such as:
Signs of dissatisfaction or risk of disengagement
Loyalty and membership-related intent
Opportunities for cross-sell or upsell
Unmet expectations expressed during support interactions
Because these signals come directly from live conversations, they often appear earlier than transactional data or survey results.
This allows Mars to:
Identify customers at risk before churn shows up in downstream metrics
Spot missed upsell opportunities embedded in support conversations
Align customer experience and marketing efforts around the same signals
Use conversation-level insight as an input into personalization and retention strategies
Customer language becomes a forward-looking input into growth and loyalty decisions, not just a retrospective explanation.
Use Case 03
Product and Experience Feedback at Scale
“It allows teams across the organization to work from the same data without rebuilding logic or relying on manual effort.” — Digital Experience & Analytics Leader, MARS Petcare
Mars receives large volumes of open-ended feedback through chatbot conversations and surveys. Historically, this feedback was reviewed manually, summarized in decks, or sampled sporadically, making it difficult to spot trends early or compare insights across brands.
Dimension Labs transforms this free-text feedback into structured signals that can be analyzed alongside other business data.
At a high level, Mars structures feedback around:
Product and brand mentions
Common themes and topics
Experience friction and confusion points
Sentiment and emotional tone
This enables Mars to:
Quantify feedback that was previously anecdotal
Track recurring product and experience issues across brands
Compare feedback trends over time, including before and after launches or rebrands
Ground product and content decisions in actual customer language at scale
Rather than relying on word clouds or summaries, teams work with structured feedback data that reflects the full population of customer conversations.
Conclusion
Conclusion
Mars Petcare generates millions of customer conversations across brands, channels, and markets. By establishing Dimension Labs as the Meaning Layer for customer language, the company created a durable foundation for turning raw conversations into structured, analytics-ready signals.
As digital engagement continues to scale, this Meaning Layer ensures customer language remains consistently understood, comparable over time, and ready to drive future improvements across products, experiences, and growth.
DISCOVER MORE
