# Managing Shifting Customer Expectations With a Diversified Investment Strategy
This is a framework for technology B2C teams running SAFe with 6–15 people who are struggling to keep up with multiplying customer demands. It borrows a portfolio allocation model from Li Ka-shing's approach to building CK Hutchison and applies it directly to how product teams choose what to build.
---
## The Problem You Probably Recognize (1/30)
. You delivered group challenges and leaderboards. Satisfaction went up again.
Last month, everything changed. Users now want AI-powered personalization. They also want offline mode. They also want Apple Watch integration. They also want meal planning. Your product manager picks what feels most important, but there is no data behind the picks. The team builds features some users want but not all users need. Satisfaction scores are flat. Churn is up four percent this quarter. (3/30)
You cannot build everything. You need a framework for choosing. That framework is an expectation portfolio.
---
## The Core Principle
Li Ka-shing never put all his capital into one sector. He spread investments across ports, retail, telecommunications, energy, and infrastructure. More importantly, he allocated based on growth potential. High-growth sectors got more capital. Low-growth sectors got less. He reviewed the data continuously and shifted money fast. (4/30)
Your team faces the same allocation problem. You cannot build every feature customers want. You must allocate capacity based on growth potential. Growing expectations get more capacity. Shrinking expectations get less. The allocation must be data-driven and reviewed every sprint. Create an expectation portfolio. Measure the growth rate of each expectation. Allocate capacity based on growth rate. Review and adjust every sprint.
---
## Step 1: Build an Expectation Portfolio (5/30)
Li tracked every investment in a portfolio with sector, amount, growth rate, and risk. You should do the same with customer expectations.
Your product manager creates a simple spreadsheet with five columns: expectation, volume, growth rate, current investment, and recommended investment. Populate it using three data sources: app store reviews scraped weekly by your data analyst, in-app feedback categorized by topic, and support tickets tagged by topic. (6/30)
- Social features. 1,200 requests. Growth down 12%. Current investment 15%. Recommended 5%.
- Progress analytics. 800 requests. Growth down 3%. Current investment 10%. Recommended 5%.
The full picture is now visible. The team can see which expectations are growing, which are shrinking, and where current investment does not match reality. (9/30)
For a SAFe team of six to fifteen people, create this portfolio before your first program increment. It should take about four hours and pull from at least three data sources. Treat it as a planning input for PI planning.
---
## Step 2: Review Every Sprint and Shift Capacity
Li reviewed his portfolio daily and shifted capital as soon as growth rates changed. Your team should do the same every sprint. (10/30)
Sprint two to sprint three, offline mode requests rise from 2,600 to 3,100. Growth goes from 12% to 18%. The allocation increases from 15% to 20%. Workout variety requests drop from 2,300 to 1,900. The allocation decreases from 15% to 10%. The team shifts five percent of capacity from workout variety to offline mode.
The review is fast, data-driven, and happens every sprint. The team is always investing in the fastest-growing expectations. (14/30)
The review should take no more than thirty minutes and result in at least one capacity shift. Make it part of your sprint review.
---
## Step 3: Write a Thesis for Every Expectation
Li treated every investment as a thesis. He had a hypothesis for why it would grow and a kill switch for when to exit. Your team should do the same. (15/30)
Apple Watch integration. Hypothesis: Thirty-eight percent of our users own an Apple Watch. If we integrate, engagement among Apple Watch owners will increase by at least 20%. Kill switch: If integration does not increase engagement among Watch owners by at least ten percent after two months, we stop.
These theses give the team discipline. Everyone knows when to stop. The kill switches prevent throwing good effort after bad ideas. (19/30)
Every expectation in your portfolio should have a thesis before you start building it. Write them as part of PI planning. Treat them as planning artifacts.
---
## Step 4: Diversify, Do Not Go All In
Li never put all his capital into one sector, even when one was growing fast. He kept investments spread across multiple sectors so that a decline in one would not sink the whole portfolio. Your team should do the same with capacity. (20/30)
The product manager allocates capacity across at least four expectations based on the portfolio:
- AI personalization: 35% (four people)
- Offline mode: 20% (two people)
- Apple Watch integration: 15% (two people)
- Meal planning: 10% (one person)
The remaining 20% is reserved for maintenance, bug fixes, and technical debt. (21/30)
No single expectation gets more than thirty-five percent of capacity. This matters. If the recommendation engine does not work out, the team still has three other expectations in progress. The team is not dependent on one bet. Running four efforts in parallel also means you learn fast across multiple areas. That learning informs your next allocation decisions.
Diversified capacity allocation should be a decision made during PI planning.
---
## Step 5: Rebalance Quarterly (22/30)
Li ran quarterly portfolio reviews that were more detailed than his daily monitoring. He checked whether each investment thesis was still valid. Failed investments were exited. Validated winners got more capital.
Your team should run the same review every quarter. Examine every expectation in the portfolio and ask three questions:
Did the expectation meet its kill switch criteria? If yes, exit it and reallocate the capacity. (23/30)
Did the expectation exceed its hypothesis? If yes, double down and increase allocation.
Are there new expectations that should be added? If yes, add them and assign a starting allocation. (24/30)
. A new expectation, voice-guided workouts, shows 1,500 requests at 22% growth. It gets added with a 10% allocation.
Every quarter, exit what isn't working, scale what is, and test what is new.
The rebalancing should take about two hours, result in at least one exit and one new addition, and happen during PI planning.
---
## Stop Guessing, Start Allocating (26/30)
Churn was up four percent last quarter because the product manager was guessing. Turn that around by allocating like an investor. An 18% retention increase within one quarter is realistic when you make decisions based on data instead of instinct.
The expectations will keep multiplying. That is fine. You do not need to build everything. You just need a system that tells you what to build next. (29/30)
agile