Netflix 2011: reasons nobody could look up
What is not open inside cannot be explained outside.
In the summer of 2011 Netflix separated streaming and DVD by mail: first in price, then as a brand of its own. The direction was right. Even so, Netflix lost 805,000 subscribers in one quarter, and the share price fell by almost 80 percent. To the outside, the reasons and criteria of the decision were not explained. Presumably they were not in the open inside the company either: the managers kept their doubts to themselves. A decision that is not transparent inside cannot be explained outside.
Two businesses under one price
Until July 2011 the bundle of unlimited streaming and one DVD by mail cost 9.99 dollars a month. On 12 July Netflix announced that the two would be charged separately: 7.99 dollars for streaming, 7.99 dollars for DVDs, 15.98 dollars together [1]. Anyone who wanted to keep both paid 60 percent more, existing members from 1 September [2].
The stated reason was that prices should better reflect the costs of each offer and give members a choice [2]. The strongest driver was not in it. Rights for streaming content were getting expensive: for 2012 Netflix planned almost double the spending of 2011 [5]. On 1 September the channel Starz broke off talks about renewing its film rights [8].
Behind the split there was also a conviction held by chief executive Reed Hastings. The future lay in streaming, and companies that are big in one business often miss the next. As examples he later named AOL's dial-up access and the bookstore chain Borders [3].
Qwikster: announced and withdrawn after 22 days
On 18 September Hastings wrote a letter to the customers. It begins with an admission: “I messed up. I owe everyone an explanation.” [4] Hastings conceded that he should have explained the price change himself, and supplied the reasons [3].
In the same letter he announced the next step. DVD by mail was to become a brand of its own under the name Qwikster: its own website, its own billing, its own queues and ratings, none of it connected to Netflix [4]. For about 11.6 million customers who used both services, that meant two accounts [5].
On 10 October Netflix withdrew the decision. The reason given was that two websites would make things harder for many members. It stayed at one website, one account and one password [6]. The higher prices stayed.
The decision as a map
The map shows the second decision, the Qwikster brand of 18 September. In one future DVD by mail becomes a service of its own. In the other it stays one service with two plans, as it had been since July.
The first state holds what was put forward as the reasoning. The second adds two rows: the objections with which Netflix withdrew the decision 22 days later. By Hastings's own account many managers had severe doubts and kept them to themselves [9][10].
F1: DVD by mail becomes a service of its own with its own website and billing. F0: one service, one website, two plans as since July.
What was on the table
The chief executive's reasons and, as the only objection, the effort. This is how the decision looked as long as nobody objected.
- F1 · SPLIT
- Streaming grows faster as a business of its own 69%
- DVD by mail gets its own management and its own offer 19%
- F0 · ONE SERVICE
- No second website, no second billing 12%
Size growth of a service with 21.45 million members
How it is scored
- V 9
- Netflix sees its future in streaming. Only the trust of its customers weighs as much.
- I 6
- With its own management and marketing the streaming service wins new members faster. That would be a lot, but it concerns only the growth and not the 21.45 million members who are already there.
- P 4
- The mechanism is plausible but not supported by evidence. The plans had been separate since July, so the costs were already assigned. No figure shows what a separate brand adds beyond that.
Size 13.93 million DVD members, now with video games
How it is scored
- V 5
- Running a shrinking business well to its end counts, but less than growth.
- I 2
- DVD customers get video games as an add-on, and the mail business decides faster about itself. This happens in a business Netflix itself considers to be winding down.
- P 6
- A separate chief executive and video games have been announced. Whether DVD customers want that is open.
Size one-off effort for the rebuild and operations
How it is scored
- V 4
- Internal effort that does not touch the customers.
- I 1
- Netflix saves building and running a second website and a second billing system. These are internal costs, small against everything else on the table.
- P 9
- Certain: a company that does not split builds nothing new.
With the unvoiced objections
Two rows are added. Both were known or foreseeable on 18 September, and Netflix named both itself when it withdrew the decision.
Size growth of a service with 21.45 million members
How it is scored
- V 9
- Netflix sees its future in streaming. Only the trust of its customers weighs as much.
- I 6
- With its own management and marketing the streaming service wins new members faster. That would be a lot, but it concerns only the growth and not the 21.45 million members who are already there.
- P 4
- The mechanism is plausible but not supported by evidence. The plans had been separate since July, so the costs were already assigned. No figure shows what a separate brand adds beyond that.
Size 13.93 million DVD members, now with video games
How it is scored
- V 5
- Running a shrinking business well to its end counts, but less than growth.
- I 2
- DVD customers get video games as an add-on, and the mail business decides faster about itself. This happens in a business Netflix itself considers to be winding down.
- P 6
- A separate chief executive and video games have been announced. Whether DVD customers want that is open.
Size one-off effort for the rebuild and operations
How it is scored
- V 4
- Internal effort that does not touch the customers.
- I 1
- Netflix saves building and running a second website and a second billing system. These are internal costs, small against everything else on the table.
- P 9
- Certain: a company that does not split builds nothing new.
Size about 11.6 million customers with two websites, two charges, two queues
How it is scored
- V 8
- Simplicity for the customer is what made Netflix big.
- I 7
- About 11.6 million customers, almost half of 23.79 million, will log in to two websites, see two charges and keep two queues. For them the service becomes more cumbersome every month.
- P 9
- Follows from the announcement itself: the two websites are not connected, and a rating counts on only one of them.
Size the customer base and the reputation that holds it
How it is scored
- V 9
- Customer trust: the good every subscription lives on.
- I 10
- Customers cancel, and with every cancellation a payment is missing month after month. Those who leave in anger advise others against the service, and reputation decides about new customers. This can affect the whole base of 23.79 million. That is the largest difference on the table, which is why it stands at 10.
- P 5
- Since 12 July the reaction to a change without explanation had been on record. That a second intervention would add to it is likely. How much could not be quantified on 18 September.
Decision map: the second state has two more rows. The numbers in the other rows are the same.
How to read the map
Each row names exactly one impact: one difference between the two futures. V is the weight of the good at stake (0 to 10), I is the size of the difference (10 is the largest on the table, all others are in proportion to it), P is the plausibility given the evidence on the day of the decision (0 to 10). The numbers are assumptions of this reconstruction, not measurements. Each one carries its reason so that it can be challenged.
The scoring uses what was known on the day of the decision. What became known later is in the last section and enters no number on the map.
Not open inside, not explainable outside
- 12 July
Streaming and DVD are charged separately: 15.98 instead of 9.99 dollars a month.
Reason only hinted at - 1 September
The new prices apply to existing members. The channel Starz breaks off talks about film rights.
Cost pressure becomes visible from outside - 18 September
Hastings apologizes, explains the split and announces Qwikster.
Reasons supplied late, doubts among managers unvoiced - 10 October
Netflix withdraws Qwikster: one website, one account, one password.
Customers' objection adopted - 24 October
Letter to shareholders: 805,000 fewer subscribers. Streaming content will cost almost twice as much in 2012.
Cost reason quantified for the first time
Fig. 1: The cost reason for the price split was quantified 104 days after the announcement. Sources: Netflix announcements and the Q3 2011 letter to shareholders.
Inside the company the decision was not in the open. Hastings said in 2020 that his leading people had been too deferential: he had been right so often that they assumed he saw something they did not [9]. Many had severe doubts and did not know that others shared them [10]. The reasons for Qwikster thus stood unexamined in one person's head. By Hastings's account, Netflix introduced a shared record of reasons and objections only afterwards [9].
To the outside, the reason was missing. On 12 July customers saw a price that rose by 60 percent and a reason given in half a sentence. That streaming content would cost almost twice as much the following year they learned in October. Netflix later wrote itself that the lack of explanation had compounded the problem [5].
The two are presumably connected. A company that has not written out the reasons and criteria of a decision internally and tested them against objections has nothing finished to put in front of its customers either. The explanation then comes after the fact, as in the letter of 18 September. For Netflix this connection is not documented. It is a reading that fits both gaps.
What would have changed the decision
Netflix apparently lacked a procedure in which reasons and objections are written down and tested before a decision is announced. What such a procedure would have changed shows in three places.
First, inside the company. A map requires the other side to be filled in. An empty column stands out where silence in the room otherwise counts as agreement. Whoever enters an objection contradicts a row and not the chief executive. The two objections with which Netflix withdrew Qwikster after 22 days would have been on the table before the announcement. With them the split is clearly behind.
Second, the map separates two decisions that appeared as one in the summer of 2011. The price split had a strong reason with high plausibility: rising content costs, documented in the company's own plans. The Qwikster brand rested on a mechanism without evidence and had two heavy objections against it. On two maps it would have been visible that the first decision holds and the second does not.
Third, to the outside. The announcement to customers can be written from a completed map, because the reason is there with its number. A customer who reads that film rights will cost almost twice as much next year does not have to like the higher price. She will not take it for arbitrariness, though.
What that would have been worth can be narrowed down but not quantified. A price 60 percent higher would have caused cancellations even with an explanation. But Netflix wrote itself that the lack of explanation had compounded the problem [5]. What would have been avoided is the announcement and withdrawal of a brand within 22 days, and with it the second loss of trust in ten weeks.
Netflix drew a related lesson from the case. Since then, for major decisions, every manager writes a number from minus 10 to plus 10 and the reason for it into a shared spreadsheet [9]. Hastings later said that for Qwikster many values around minus 7 would have appeared there [10]. The decision map goes one step further: it asks for value, size and evidence per argument. That shows what two assessments actually disagree about.
What happened afterwards
At the end of 2011 Netflix again had 24.4 million subscribers in the US [11]. The price split stayed, and the direction held. The number of DVD members already fell in the fourth quarter of 2011, from 13.93 to 11.17 million [11]. Netflix shipped its last DVDs on 29 September 2023 [12].
In October 2011 Netflix wrote to its shareholders that what it had misjudged was how quickly to move toward streaming [5]. The share price stood at 62.37 dollars at its low in the fourth quarter, after 304.79 dollars at the high in the third [7].
What is not open inside cannot be explained outside.
A decision is announced, the reasons follow only after the protest, and in the end it is withdrawn. This keeps happening, even to companies that know their customers well.
- 2023 to 2024
Unity
In September 2023 Unity announced a fee per installation of a game. Ten days later the company apologized: it should have talked more with its customers before the announcement. In October the chief executive stepped down, and a year later the fee for games was dropped entirely [13][14][15][16].
- 1985
Coca-Cola
In April 1985 Coca-Cola replaced its formula with one that had done better in taste tests with nearly 200,000 people. The tests did not show what tied customers to the old formula. After 79 days it came back [17].
- 2010
Gap
In October 2010 Gap introduced a new logo and explained it only after the protest. A few days later the company returned to the old logo. By its own account it had involved its customers too late [18][19].
Sources
- Netflix: press release of 12 July 2011 on the separate plans
- Engadget (12 July 2011): wording of the stated reason and effective date for existing members
- ABC News (21 September 2011): Netflix CEO apologizes and announces Qwikster
- Reed Hastings: An Explanation and Some Reflections, 18 September 2011 (reprinted by Gaming Nexus)
- Netflix: Q3 2011 letter to shareholders (24 October 2011)
- Macworld (10 October 2011): Netflix reverses Qwikster
- Netflix: annual report (10-K) for 2011
- Associated Press (2 September 2011): Starz ends its talks with Netflix
- Reed Hastings on the a16z podcast (15 September 2020): Designing a Culture of Reinvention
- CNBC (20 April 2024): how Reed Hastings makes big decisions at Netflix
- Netflix: Q4 2011 letter to shareholders
- Netflix (18 April 2023): the end of DVD by mail
- Unity (12 September 2023): Unity plan pricing and packaging updates (archived copy of 14 September 2023)
- Unity, Marc Whitten (22 September 2023): An open letter to our community (archived copy)
- Unity (9 October 2023): Unity Announces Leadership Transition (exhibit to Form 8-K)
- Unity, Matt Bromberg (12 September 2024): Unity is canceling the Runtime Fee
- The Coca-Cola Company: New Coke, company history
- NPR (8 October 2010): The Gap Gets Backlash For Logo Makeover
- NPR (12 October 2010): Gap Drops New Logo
Decision under uncertainty Challenger, 1986
A risk without a number is not a risk of zero.
The engineers could not quantify the risk. The map shows they did not have to: it was enough to know on which side of the tipping point it lay.
Read the case →
Decisions over time Kodak, 1981 to 2003
A good decision has an expiry date.
With what was known in 1981, staying with film was right. The reasoning had an expiry date. When it was reached, the decision stayed in force.
Read the case →
Deciding by one's own values Volkswagen, 2006 to 2015
Values count only when they are part of the decision.
The objections were on the table, the values were on paper. Under pressure of time and cost, neither counted. As a row in a decision, a value is harder to pass over in silence or to talk down.
Read the case → The method behind the mapsThe full derivation of V × I × P is on the page Reason Based Decision Making for Business.
To the derivation →