IGNITE EXTENDED / LESSON 15
What should stay consistent when a business changes its model?
A before/after strategy map and a pivot decision rule.
The Netflix history moves through physical distribution, streaming, content and international expansion. Each transition changes the customer proposition, economics and capabilities required. Calling all of them “growth” hides the work needed to make the next model function.
Map the old and proposed models side by side. Identify which capabilities transfer, which become less valuable and which must be developed. Then examine whether the transition undermines the current business before the new one can support it. Historical outcomes should not substitute for the uncertainty decision makers faced at the time.
A pivot changes an important element of the venture’s model in response to learning. A change of activity without a reasoned connection to evidence can become an unrelated expansion. State which observation rejects the old assumption and why the proposed change addresses it.
Preserve what remains supported. A team may have evidence for a customer need but weak evidence for a revenue model or channel. Separating those elements reduces the risk of discarding useful learning or carrying an unsupported assumption into the next version.
Use the archived Netflix discussion to trace a strategic transition. Each stage changes several connected activities rather than replacing only the customer-facing format.
An offer based on physical media requires inventory, delivery and a way to make the waiting period acceptable. Customer convenience depends on the complete delivery system and the pricing arrangement.
A recurring payment changes the revenue relationship and the importance of continued customer value. Analyze use, retention and service costs together instead of treating a subscription label as sufficient business logic.
Digital delivery changes access and the operating system behind it. Rights, technology and customer expectations become part of the transition. Earlier capabilities may help in some areas and constrain others.
Ask which resources, incentives and measures must change with the strategy. A transition can require running the existing model while learning how the new one works.
This is a historical teaching sequence. The next analysis asks what changed in customers, activities and economics, then distinguishes a reasoned pivot from an unrelated expansion.
A pivot revises an important part of the business in response to what has been learned. An adjacent idea becomes coherent only if its customer, activities and economics connect to the proposed strategy.
Responds to a changed opportunity and explains how existing strengths help create new ones.
Adds an activity without a clear account of customer value, capabilities or business logic.
Illustrative application: a team learns that supervisors value an audit trail more than faster messaging and changes the offer accordingly. Adding an unrelated consumer app because it sounds promising does not follow from that evidence. Spell out what is retained, what changes and why.
The strategy memo should connect new evidence to a revised choice and the capabilities that choice requires.
A strategy requires particular tasks and behaviors. People, formal structure, rewards and cultural expectations influence whether those behaviors occur. A company can describe a new strategy while continuing to reward the choices that supported the old one.
The alignment model asks how the organization must change to execute the proposed strategy. Identify a concrete contradiction and its operating consequence. A consistent organization does not eliminate uncertainty, but an inconsistent one can prevent the strategy from receiving a fair test.
A recommendation under uncertainty should identify the evidence that would lead to revision. This protects a strategic direction from changing with every anecdote while also preventing commitment from becoming immune to evidence.
Specify what the next period of work should establish, when the decision will be reviewed and which alternatives remain available. Consider the cost of waiting and the cost of reversing the choice. An experiment that arrives after an irreversible commitment may have little decision value.
A change in strategy can require changes in how the organization works.
A vision describes a desired direction. Evidence helps choose which actions are plausible ways to move toward it and which assumptions need revision.
A direction that helps people understand what the organization is trying to achieve.
Information about whether the chosen path can work. Vision can guide choices without making an unsupported path true.
Illustrative application: “make handoffs dependable” can remain the objective while repeated tests reject a notification-heavy product. The team might try an authoritative record or service process instead. Preserving the exact feature despite contradictory observations confuses the means with the objective.
State which commitments define the purpose and which are hypotheses about how to accomplish it.
Click either page to read it at full size. In the reader, use ← → to switch between these two pages.
Click either page to read it at full size. In the reader, use ← → to switch between these two pages.
Selected reading: PDF pages 2, 3, 4, 5
The opening excerpt frames Netflix’s history as changing business models and describes early subscription, delivery and recommendation choices. It also presents a deliberately designed approach to organizational culture.
The early move from per-rental transactions to subscription changes the customer relationship and the revenue logic.
Physical media, packaging and return logistics affect whether a mail-based offer can deliver convenience at viable cost.
The excerpt introduces recommendations alongside inventory management. A customer-facing capability belongs to a wider operating system.
The culture discussion describes high expectations and substantial discretion. The case links management choices to the work the company expects to perform.
Selected reading: PDF pages 2, 3, 4, 5
These pages introduce later eras but mainly describe early choices. They do not establish every causal explanation for the later streaming and content businesses.
Changing a price or delivery format affects costs, behavior and required capabilities. Examine the connected design before calling a transition a simple product switch.
A policy drawn from one company can have different effects under different work and labor conditions. Evaluate the mechanism and trade-offs before copying it.
The teaching materials examine how logistics, customer knowledge, access, content and organizational choices mattered differently across historical eras. A capability useful in DVD delivery was not automatically an advantage in streaming or content.
How should a firm decide which strengths to preserve and which to rebuild?
Netflix launched an ad-supported plan at $6.99 a month on November 3, 2022, in nine countries including the US. Selling advertising requires capabilities the subscription model did not: ad sales, measurement and an ad-tech stack.
In May 2023 Netflix rolled out paid sharing in the US, charging $7.99 a month for each extra member outside the household. By its own estimate, more than 100 million households had been sharing accounts they did not pay for.
In April 2024 Netflix said it would stop reporting quarterly subscriber numbers from Q1 2025 and steer investors to revenue and profit. Once revenue per member varies by tier and ads, a subscriber count no longer measures the business, so the measures had to change with the strategy.
Sources: TechCrunch: Netflix's ad-supported tier set to launch on November 3 · CNBC: Netflix password-sharing crackdown rolls out in the US · Gulf News: Netflix to hide its subscriber numbers starting in 2025
Fictional firm: an established physical service has loyal customers and delivery expertise. A digital alternative removes shipping but requires new technology and rights. Current customers have shown interest, but repeated use is untested.
Compare staying, a staged test and a full pivot. Map the capability gaps and state a trigger for the next commitment.
Fourteen months after launch, Relay's data shows that multi-location PT groups retain about 3x longer than independent clinics and buy 10 or more locations at once. Independents churn faster. Daniel wants to launch Relay Scheduler, an AI assistant on a third-party LLM API that converses with patients, proposes slots and predicts who will accept. LLM inference would lower gross margin from about 80% to 72% unless it is managed. Priya Raman, Head of Growth since month 9, runs a founder-style direct sales motion built for single clinics.
Write a before/after strategy map (customer, offer, activities, economics, organization), a decision rule that says when to commit, stage or stop, and an alignment plan naming three organizational contradictions and how to fix them.
'AI scheduling is the future' is a vision. The pivot is justified by retention and deal-size data for groups. Write down which observation rejects the old segment and which assumption the new model still has to prove.
Moving to Scheduler adds LLM inference cost (80% to 72% gross margin) and longer enterprise sales cycles. Recompute LTV and CAC under the new model instead of carrying v1's numbers forward.
If sales pay, onboarding processes and engineering's definition of done still serve single clinics, the pivot will fail in execution and be judged a strategic mistake.
Relay's before/after strategy map, a staged pivot decision rule with three thresholds and a review date, and an alignment plan covering sales pay, onboarding and engineering's definition of done.
Next lesson: Financial statements & economic reality
Check your understanding
Question 1
Relay's data shows groups retain about 3x longer than independents. Which proposed change is a coherent pivot rather than an unrelated expansion?
The evidence concerns who retains and buys at scale, so changing segment and the activities that serve it follows directly. The dental version might later be adjacent, but nothing in the retention data supports it now.
Question 2
Illustrative figures: independents churn 4.5% a month and groups 1.5%. At $399 a location, 80% margin for independents and 72% (with LLM costs) for groups, what is the approximate LTV per group location?
Lifetime is 1/0.015, about 67 months, and 399 x 0.72 x 67 = about $19,250. $21,400 uses the old 80% margin and ignores the inference cost that comes with Scheduler; $10,500 is the blended v1 figure.
Question 3
Netflix's DVD-era advantages included envelope patents, warehouse optimization and shipping operations. What happened to these when Netflix moved to streaming?
Logistics capabilities have no role in digital delivery. What carried over were customer-side assets. The tempting answer, 'barriers', assumes an advantage keeps working once the competitive setting changes.
Question 4
Relay announces a security-first enterprise strategy, but engineers are still promoted for feature velocity and sales reps are paid per new logo. What is the most likely result?
Incentives, promotions and processes shape behavior more than announcements do. Unless the levers change, the old model keeps being executed and the pivot may be rejected for the wrong reason.
Question 5
Which decision rule is best for Relay's pivot?
A good rule names the evidence, the thresholds, the review date and the alternatives, and it comes before an irreversible commitment. A full pivot now treats the vision as proof; waiting on independents or rivals tests the wrong thing.
Question 6
A common misconception: 'A pivot means abandoning the original idea.' Which statement is more accurate?
Relay keeps its purpose (filling empty slots) and its validated fill-rate result, and changes the segment, offer and go-to-market. Throwing out supported learning is as costly as keeping an unsupported assumption.
Workbook
Module output: A business-model map, market/game models and an advantage-and-pivot memo.
Library
8 appear in the slides with notes; the rest are kept here for reference.
Source S25 · Feinberg · 40 pages
Source S26 · Feinberg · 23 pages
Source S02 · Feinberg · 15 pages
