agile# How to Use the Platform Business Model to Balance Innovation with Maintenance in Entertainment B2B (1/27)
An entertainment B2B family business running Crystal has a team of six to fifteen people and a problem that sounds familiar. The company provides a white label ticketing platform for live event venues. Seat selection, payment processing, barcode generation, box office management. Eight years in business, twenty nine employees, eleven in product development. One team. And that team is stuck. (2/27)
The stuckness lives between innovation and maintenance. New features on one side. Keeping the existing platform stable on the other. The platform carries significant technical debt, which means every new feature takes roughly twice as long as it should. That slowdown kills the team's ability to innovate. The company is falling behind competitors. Clients are leaving. Last quarter, lost clients cost the company $71,000. That was 26% of quarterly revenue. (3/27)
This is the kind of problem that breaks small teams. But it's not a new problem. Jack Ma faced something similar when he built Alibaba. His solution was the platform business model, and it maps directly onto what Crystal needs right now.
## What Ma Actually Did (4/27)
Ma saw the same tension at Alibaba. Build new things or maintain existing things. Most companies treat it as a choice. Pick one, sacrifice the other, hope nothing breaks. Ma thought that framing was wrong. The tension between innovation and maintenance is a false choice. Acting on it creates stagnation. Stagnation kills companies. (5/27)
So he built a platform. The principle was straightforward: build a foundation that others can build on. A stable core that rarely changes. Modular extensions that snap on without touching existing code. Maintenance stays low because the core is solid. Innovation stays high because new features are just plugins. The team stops choosing and starts building. (6/27)
Ma applied the same logic to team structure. He didn't split the team. He split the platform. Core team maintains the foundation. Extension team builds new features. Separation of concerns creates balance.
For Crystal, the situation is the same. The team is stuck. Technical debt makes innovation slow. Falling behind costs real money. Ma's answer is the same answer: build a platform that others can build on. The platform creates the balance. The balance eliminates the tension. (7/27)
## The Core Principle
The best way to balance innovation with maintenance is to stop treating them as competing priorities. Build a platform architecture where a stable core minimizes maintenance and modular extensions maximize innovation. The team is no longer choosing between new features and technical debt. They are building a foundation that makes both possible. (8/27)
Ma didn't balance these at Alibaba by forcing the team to pick a side. He built a platform with a stable core and modular extensions. The core minimized maintenance. The extensions maximized innovation. The architecture eliminated the tension. (9/27)
Crystal's team is stuck in the same trap. Technical debt slows innovation. Falling behind costs $71,000 a quarter. The platform business model says the same thing it said to Ma: build a platform that others can build on. That platform creates balance.
## Four Steps to Apply the Platform Business Model
1. Separate the Platform into a Stable Core and Modular Extensions (10/27)
Ma split Alibaba's platform into two layers. The core was maintained. The extensions were innovated. Both happened in parallel. That parallelism created balance.
Crystal should do the same. The tech lead separates the platform through an architecture refactor. It takes about three weeks. The result is two layers. (11/27)
Layer one is the stable core. This is the ticketing engine. Seat selection, payment processing, barcode generation, box office management. It is stable by design. It does not change often. That stability minimizes maintenance and frees up capacity for innovation. (12/27)
Layer two is modular extensions. New features are built as plugins. They connect to the core through APIs. They do not touch existing code. That means innovation does not create new technical debt. Maintenance stays low. Innovation stays high. (13/27)
When one team completed this separation last month, the results were immediate. Six team members maintained the core and fixed 23 technical debt items. Bug count dropped 41%. Five team members built extensions and delivered four new features. Client retention increased 18%, saving $34,000 in revenue that would have walked out the door. (14/27)
For Crystal's team, the separation should take no more than three weeks. It should be a reflection workshop output. Two layers. Stable core and modular extensions. That is the starting point.
2. Allocate a Fixed Percentage of Every Iteration to Core Maintenance
Ma dedicated 20% of every sprint at Alibaba to core maintenance. Technical debt was addressed continuously. It never accumulated. Innovation was never slowed by a growing backlog of old problems. (15/27)
Crystal should allocate a fixed percentage of every iteration to core maintenance. Twenty five percent is a strong starting point. Every iteration then has two types of work.
Core maintenance gets 25% of capacity. That time is dedicated to technical debt. Fixing it continuously prevents accumulation. Extension innovation gets 75% of capacity. New features are built as modules and added to the core without creating new debt. (16/27)
Last iteration, one team applied this split. They spent one week on technical debt and fixed eight items. Bug count dropped 15%. Support tickets went down. Client satisfaction went up. Churn went down. That saved $12,000.
For Crystal, the fixed percentage should be at least 20%. It should be applied every iteration without exception. This is a planning decision, not a nice-to-have.
3. Build an Extension Marketplace Where New Features Are Plugins (17/27)
Ma built an extension marketplace at Alibaba. A catalog of plugins that could be added or removed without touching the core. That created flexibility. Flexibility created speed. (18/27)
Crystal should build the same thing. A catalog of plugins. Each plugin is a new feature built as a module with an API interface. Plugins connect to the core without touching existing code. They can be added or removed without risk. That lets the team experiment freely. Experimentation drives innovation. Innovation drives competitive advantage. (19/27)
One team built their extension marketplace in two weeks. The catalog had six plugins: a dynamic pricing engine that adjusts ticket prices based on demand, a waitlist management system for sold out events, a group booking tool, an accessibility seating module, a merchandise upsell feature, and a real time analytics dashboard.
Clients could choose which plugins they wanted. That customization increased satisfaction. Higher satisfaction reduced churn. Churn reduction saved $25,000. (20/27)
For Crystal, the marketplace should have at least five plugins. Each one should be addable and removable without touching the core. This is a reflection workshop output.
4. Run a Feedback Loop Every Iteration to Measure the Balance
Ma ran a feedback loop at Alibaba to measure the balance between core health and extension delivery. The team could adjust when one side started to slip. Neither was neglected. (21/27)
Crystal should run the same loop. A thirty minute meeting at the end of every iteration. Two metrics on the table.
The core health score is a number from one to ten. It is based on three sub metrics: bug count, unresolved technical debt items, and test coverage. The extension delivery rate is the count of plugins delivered per iteration. Together they tell the team whether the balance is holding. (22/27)
Last iteration, one team ran the loop and found their core health score at six. Below the target of eight. Technical debt was accumulating. The fix was to increase the maintenance allocation from 25% to 30%. Extension delivery dropped to 70%. That tradeoff was worth it. Core health improved to eight. Bug count dropped 22%. Support tickets fell. Client satisfaction rose. Churn fell. Savings: $14,000. (23/27)
For Crystal, the feedback loop should happen at the end of every iteration. At least two metrics. At least one adjustment. This is a reflection activity.
## Closing on Platform Over False Choice (24/27)
Jack Ma did not build Alibaba by forcing his team to choose between innovation and maintenance. He built it by separating the platform into a stable core and modular extensions. By allocating a fixed percentage of every iteration to core maintenance. By building an extension marketplace where new features were plugins. By running a feedback loop every iteration to keep the balance. (25/27)
Crystal's twenty nine employee company is losing $71,000 a quarter because the team is stuck in a false choice. The platform business model eliminates that choice.
Separate the platform into a stable core and modular extensions this iteration. Allocate 25% of next iteration to core maintenance. Build the extension marketplace. Run the feedback loop at the end of the iteration.
The team stops choosing. The platform does the balancing. The $71,000 stops walking out the door. (26/27)