← all stories

After the Web · Era 3 · The early Web

1997–2010public company (IPO 2002, Nasdaq: NFLX); streaming since 2007

Netflix

Netflix began as a DVD-by-mail rental site in California (1998), switched in 1999 to a subscription with no late fees, and by 2002 had gone public with half a million subscribers. According to Randolph, in 2000 the founders tried without success to sell the company to the Blockbuster chain for $50 million; by the end of 2007 Netflix had 7.5 million subscribers and movie viewing over the internet, and in 2010 Blockbuster went bankrupt.

Founders Reed Hastings · Marc Randolph
Domains netflix.com
subscriptione-commerceecommerce-logistics

In brief

Netflix is a company that in 1998 began sending DVDs by mail, and in September 1999 switched to a subscription with no late fees and no due dates. By its IPO in May 2002 it had more than 500,000 subscribers and $75.9 million in revenue for 2001, with a loss of $38.3 million. In 2000 the founders, according to Randolph, tried to sell the company to Blockbuster for $50 million and were turned down; Blockbuster went bankrupt in September 2010. The story is useful because the subscription here is not "a way to take money" but a reworking of the product itself: what customers used to be punished for was taken out of it.

How it started (the founders)

Reed Hastings founded Pure Atria Software in 1991 and sold it to Rational Software in August 1997; from October 1996 Marc Randolph was vice president of marketing at IntegrityQA and its successor, Pure Atria. Netflix was incorporated in Delaware in August 1997; Randolph became its first CEO and ran the company until September 1998, when Hastings became CEO. According to Randolph's account as retold by CNBC, the idea grew out of a search for something like Amazon.com, but for a different product, and the test was sending a compact disc by mail: it arrived intact estimate.

The site went live in April 1998 and at first rented out DVDs one at a time: in January 1999 a rental cost $4.00 and a "Reserve Rental" $5.00, and the catalog held more than 2,600 discs. Discs were sold as well; sales of new DVDs were stopped in March 1999. A disc had to be sent back within 7 days, otherwise the rental was extended; the extension charge (the "Extended Rental Fee") equaled the disc's suggested retail price, after which the disc stayed with the customer estimate.

Year-by-year timeline

Lesser-known but significant facts

  1. The subscription came after per-disc rental, not in its place from the very start. Until September 1999 Netflix made its money on individual rentals and DVD sales; sales were stopped half a year before the subscription.
  2. The subscription terms changed twice in 13 months. $15.95 with four movies a month (September 1999) → $19.95 unlimited (February 2000) → $19.95 with a limit of three discs (October 2000); for some customers the price went up by $4, and the S-1 calls this part of the growth in subscription revenue in 2000–2001.
  3. The subscriber acquisition cost fell 3.5-fold in three years: $110.79 (1999) → $31.39 (2002), while trial subscriptions grew from 127,000 to 1,140,000 a year.
  4. Subscriber churn was among the risk factors as early as the S-1. There the company wrote that it had to reduce the share of subscribers who leave while adding new ones; in the 10-K for 2007 average monthly churn is 4.3% (definition: cancellations in a quarter divided by the sum of subscribers at the start of the quarter and new subscribers, and then by three months), that is, about four subscribers out of a hundred every month.
  5. The densest base is where the mail is fast. In the 2002 S-1: in the San Francisco Bay Area, where the mail delivers a disc from the distribution center in San Jose in 1–2 days, more than 2.6% of households had subscribed.

Legend vs. the record

The first growth lever

The lever was not the website but a repackaging of the product: in September 1999 the per-disc charge turned into a fixed monthly fee with no due dates, and the promise of no due dates, no late fees, ever, became the slogan on the home page. The result in the S-1 figures: 107,000 subscribers at the end of 1999, 292,000 at the end of 2000, and 456,000 at the end of 2001; subscription revenue grew from $4.9 million to $35.9 million and $74.3 million.

Customers were brought in with free trial subscriptions and pay-for-performance affiliate programs. The acquisition cost kept falling: $110.79 in 1999, $49.96 in 2000, $37.16 in 2001, $31.39 in 2002. In 1999 that was almost seven months of subscription at $15.95, and in 2002 about 1.6 months at $19.95 estimate.

The paired story

The pair is Blockbuster: a rental chain that lived in part on late fees, versus a subscription with no late fees.

Parallels today (projects from the catalog)

What a builder can take from this in 2026

  1. Find what makes customers angry about paying, and remove it. For Netflix that point was due dates and late fees; by removing them, the company got its slogan.
  2. Change prices and limits in the first year, while the base is small. Two changes of terms in 13 months; over 2000 the number of subscribers grew 2.7-fold.
  3. Count acquisition cost in months of payments. At Netflix it fell from almost seven months of subscription in 1999 to 1.6 months in 2002.
  4. Do not build a strategy on one nice scene. The $50 million meeting and the Apollo 13 late fee rest on the founders' stories; the S-1 and 10-K figures can be checked, and these scenes cannot.
  5. Watch churn from your first prospectus. The risk factor about losing subscribers was in the S-1; average monthly churn was 4.3% in 2007.

Discrepancies and what we could not verify

Sources (primary first)

Internet History Keeper

Every historical dossier here is free and open, and it will stay that way. If you want the series to keep going (new dossiers, checks against primary sources, the English edition), support it with a subscription for $9.90 a month. It is support, not access, and you can cancel at any time.

Become a keeper — $9.90 a month →

The same thing, about today

The same breakdowns, but of projects launching right now: what the product is, where the first users came from, how they charge. The card is free, the full dossier is $5 (the dossier itself is written in Russian).

Browse the dossier catalog →