Direct Customer Data Flywheel
Own the relationship, learn from feedback, and build adjacent value
- Difficulty
- Advanced
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 94%
The Direct Customer Data Flywheel starts by giving customers a strong reason to transact through a channel the company controls. That preserves direct contact instead of surrendering the relationship to an intermediary. The company then captures permissioned data and makes complaints visible as operational research. Recurring problems guide improvements to the core experience, while cleaned and connected data reveals adjacent needs in the broader customer journey. New products are tested only where they add relevant value, and their usage generates further learning. Fernandez described AirAsia using low fares to drive early internet booking, retaining customer contact, displaying complaints for staff, improving internal digital processes, and exploring travel-related services such as tickets and financial services.
Origin
Tony Fernandez described AirAsia's early use of internet booking to retain customer contact, its treatment of complaints as market research, and its later effort to clean data before building a wider travel ecosystem.
Core principles
- 01Keep direct contact with the customer
- 02Use a compelling offer to change channel behavior
- 03Treat complaints as research inputs
- 04Clean data before building products on it
- 05Create adjacent value from recurring customer needs
How to run it
- 1
Create a direct-channel reason
Give customers a compelling reason to use the channel you control. AirAsia used a sufficiently attractive fare to challenge assumptions that customers would not book online.
Pro tip Lead with customer value rather than asking people to change behavior for your convenience.
Watch out Do not use coercive channel restrictions as a substitute for a better offer.
- 2
Own the relationship
Retain permissioned contact and transaction data so the company can communicate and learn directly. Reduce unnecessary dependence on intermediaries for customer understanding.
Watch out Data ownership does not remove privacy, consent, or security obligations.
- 3
Turn complaints into signals
Make customer complaints visible to the people responsible for fixing the experience. Cluster recurring issues and prioritize corrective action.
Pro tip Treat accessibility to complaints as a learning advantage, not merely a reputation problem.
Watch out A complaint board has no value if nobody closes the loop.
- 4
Clean and connect the data
Improve data quality and internal processes before attempting sophisticated analysis or new products. Reliable inputs are the base of the flywheel.
Watch out New ventures built on fragmented data can amplify bad assumptions.
- 5
Test journey adjacencies
Use observed customer needs to test services adjacent to the core transaction. Keep each experiment relevant to the customer's broader journey.
Pro tip Start with needs customers already reveal through behavior or feedback.
Watch out Customer data alone does not grant credibility in every adjacent market.
In the wild
Fernandez said many people in the region lacked internet access in 2001, but he expected an unusually low fare to motivate them to find a way to book online. The channel allowed AirAsia to keep direct contact rather than leave the relationship with travel agents.
→ AirAsia established an early direct digital relationship and retained customer data for later use.
Fernandez described displaying complaints for staff and calling them free market research. The point was not that complaints were harmless, but that they directly identified issues the airline should fix.
→ Customer criticism became a visible input to operational improvement.
Common mistakes
Collecting data without a use
Accumulation is not a flywheel. Data must improve a decision, customer experience, or relevant product test.
Ignoring consent and trust
The transcript supports retaining direct customer contact, not unrestricted exploitation. Permission, privacy, and security remain necessary constraints.
Launching adjacencies too early
Fernandez said AirAsia first worked on getting its data right. Expanding before reliable inputs risks building on noise.
Is it for you?
Best for
It is best for businesses with repeat customers, a direct digital channel, and related needs surrounding the core purchase.
Not ideal for
It is not ideal when the company lacks customer permission, reliable data, or a credible reason to enter adjacent categories.
From the transcript
“I wanted to keep the contact with my customer”
“that's free market research. That's telling you what's wrong with our airline”
“the first thing is to get the data right”
From the episode
605: He Bought an Airline for $0.30 (and made BILLIONS)