WEBVTT - Netflix's Early Years 0:00:04.120 --> 0:00:07.160 Get in touch with technology with tech Stuff from how 0:00:07.200 --> 0:00:13.840 stuff works dot com. Hey there, and welcome to tech Stuff. 0:00:13.880 --> 0:00:16.520 I'm your host, Jonathan Strickland. I'm an executive producer with 0:00:16.560 --> 0:00:19.640 how Stuff Works and I love all things tech. And 0:00:19.680 --> 0:00:26.000 we are continuing our four part series about Netflix and uh, 0:00:26.280 --> 0:00:29.520 you know. In the previous episode, I covered the founding 0:00:29.800 --> 0:00:32.600 of the company and the challenges that they face just 0:00:32.640 --> 0:00:37.200 getting it up and running, going from the ideation of 0:00:37.240 --> 0:00:40.839 what Netflix would be to building an actual working business, 0:00:41.159 --> 0:00:43.880 and today we're going to continue that story. Less than 0:00:44.240 --> 0:00:47.920 half a year after Netflix's website went live in the 0:00:47.960 --> 0:00:51.840 spring of the company was making about a hundred thousand 0:00:51.920 --> 0:00:55.760 dollars in revenue per month. Now that's not profit, that 0:00:55.880 --> 0:00:59.080 was just revenue, but it did push Netflix into being, 0:00:59.120 --> 0:01:02.240 at least in theory, a million dollar company. You know, 0:01:02.280 --> 0:01:04.880 if they bring in a hundred thousand dollars per months, 0:01:04.880 --> 0:01:07.920 you multiply that by twelve months, you realize, oh, that's 0:01:07.959 --> 0:01:10.760 one point two million dollars. That's pretty impressive. And the 0:01:10.760 --> 0:01:14.119 company wasn't even six months old yet. The model Netflix 0:01:14.240 --> 0:01:17.360 used meant customers would rent films for four or five 0:01:17.440 --> 0:01:19.880 dollars per title, and then they would keep them for 0:01:19.920 --> 0:01:22.880 a week. They could also opt to purchase a title 0:01:22.920 --> 0:01:25.760 that they liked. They could end up buying it for 0:01:25.800 --> 0:01:30.399 a thirty PC discount and just keep the DVD. The 0:01:30.400 --> 0:01:33.800 company had rented out about twenty thousand DVDs in the 0:01:33.880 --> 0:01:36.919 first four months alone, so things were going pretty well. 0:01:37.400 --> 0:01:40.360 The development team was hard at work building new features 0:01:40.400 --> 0:01:43.280 for the company. The big one was what would become 0:01:43.319 --> 0:01:48.040 called the Q as in the line, the que that 0:01:48.040 --> 0:01:51.560 would order titles customers had expressed interest in but had 0:01:51.640 --> 0:01:55.280 not yet rented, so they could create a list of 0:01:55.360 --> 0:01:59.680 titles that they were interested in and even prioritize that list, 0:01:59.720 --> 0:02:02.120 put a movies they more they wanted to see more 0:02:02.240 --> 0:02:04.200 towards the front of the queue and the ones that 0:02:04.240 --> 0:02:06.360 they wanted to see, but there wasn't as big a 0:02:06.360 --> 0:02:09.240 sense of urgency towards the back. That would be something 0:02:09.760 --> 0:02:13.120 that the development team would work on for a couple 0:02:13.160 --> 0:02:16.119 of years before it was really ready to roll out 0:02:16.160 --> 0:02:20.600 in full form, but it became a foundational element of 0:02:20.600 --> 0:02:23.480 what Netflix would be, and that The work on the 0:02:23.560 --> 0:02:25.920 queue dates all the way back to the late nineties. 0:02:26.040 --> 0:02:30.120 Right after Netflix was founded. Programmer Eric Meyer built in 0:02:30.200 --> 0:02:33.240 features in search that would lead users to movies that 0:02:33.280 --> 0:02:36.880 were in stock, and it would exclude movies that were 0:02:36.880 --> 0:02:40.239 either not yet in Netflix's library and that they didn't 0:02:40.240 --> 0:02:43.000 have a copy of that DVD yet, or they were 0:02:43.080 --> 0:02:46.839 in Netflix's library, but we're currently rented out to all customers. 0:02:46.880 --> 0:02:49.320 So that way, when you come to Netflix and you 0:02:49.360 --> 0:02:52.280 start searching around for a movie, not a specific film maybe, 0:02:52.280 --> 0:02:56.359 but maybe a movie featuring a specific actor or actress 0:02:56.480 --> 0:02:59.720 or directed by a certain person, you would get results 0:02:59.760 --> 0:03:04.080 back that only reflected available films. On the business side 0:03:04.120 --> 0:03:08.520 of things, Netflix began to feature older films and uh 0:03:08.560 --> 0:03:12.360 that was far more prevalent than new releases, and this 0:03:12.440 --> 0:03:16.080 was largely because the company wasn't able to yet negotiate 0:03:16.880 --> 0:03:21.560 with movie studios for DVD prices, and new releases tend 0:03:21.680 --> 0:03:24.760 to cost the most money. So while the prices for 0:03:24.880 --> 0:03:28.400 DVDs were significantly lower than what it would cost to 0:03:28.440 --> 0:03:31.880 buy the same title on vhs. Remember DVDs when they 0:03:31.880 --> 0:03:36.120 came out, they were priced typically between fifteen and twenty dollars. 0:03:36.280 --> 0:03:39.080 VHS tapes if you were buying a copy for a 0:03:39.120 --> 0:03:44.200 rental place like like a Blockbuster or something, those tended 0:03:44.200 --> 0:03:46.480 to be between sixty five and eighty dollars if you 0:03:46.520 --> 0:03:49.560 didn't have the leverage to ask for a lower amount 0:03:49.720 --> 0:03:54.080 and do sort of a royalties approach. So building out 0:03:54.080 --> 0:03:59.120 an inventory at scale was prohibitively expensive if you wanted 0:03:59.160 --> 0:04:02.080 to go with all new releases, so Netflix opted to 0:04:02.080 --> 0:04:06.080 go for deeper cut films that weren't typically listed at 0:04:06.120 --> 0:04:10.680 full price. Then they would run promotions on the website 0:04:10.880 --> 0:04:15.280 that would feature these deeper cuts and leverage those titles, 0:04:15.520 --> 0:04:18.320 and they would tie it into things like holidays or 0:04:18.600 --> 0:04:22.120 news events, or maybe it's an actor's birthday and they say, 0:04:22.240 --> 0:04:26.320 to celebrate so and so's birthday, why not rent this movie? 0:04:26.880 --> 0:04:31.120 And that helped keep the business going even when they 0:04:31.279 --> 0:04:35.520 weren't able to have all the latest new releases in stock. 0:04:36.080 --> 0:04:39.080 The programmers built out a feature that would group similar 0:04:39.120 --> 0:04:43.160 film buffs together in mentor groups through a process called 0:04:43.200 --> 0:04:46.719 collaborative filtering, and the idea was that these users with 0:04:46.800 --> 0:04:50.240 similar tastes might end up directing each other to new 0:04:50.279 --> 0:04:54.479 films that fit those shared interests. So you might say, oh, 0:04:54.640 --> 0:04:59.479 you really like, uh, you know, martial arts films, Well, 0:04:59.680 --> 0:05:03.200 have you seen Fatal Flying Guillotine here? You should check 0:05:03.240 --> 0:05:06.479 it out. Netflix has it that kind of thing. Netflix 0:05:06.520 --> 0:05:10.760 also began building out a larger inventory of foreign cinema 0:05:10.839 --> 0:05:16.480 titles speaking of Fatal Flying Guillotine, particularly Bollywood films, Chinese movies, 0:05:16.560 --> 0:05:20.640 Japanese movies. It became a place where you could get 0:05:20.720 --> 0:05:25.160 hold of this foreign cinema much more easily. And in California, 0:05:25.320 --> 0:05:31.320 where you had a lot of immigrants from countries like China, Japan, India, 0:05:31.680 --> 0:05:34.880 it was very popular, and of course that's where Netflix 0:05:34.920 --> 0:05:38.880 was based, so it ended up being a very uh 0:05:39.200 --> 0:05:42.920 lucrative or at least a very active part of Netflix's library. 0:05:43.320 --> 0:05:49.279 The one genre Randolph decided not to carry was hardcore pornography, 0:05:49.320 --> 0:05:51.240 and that was more out of a business concern than 0:05:51.279 --> 0:05:54.800 anything else. It wasn't that he was necessarily morally opposed 0:05:54.920 --> 0:05:58.720 to that content, but rather he was worried about it 0:05:58.800 --> 0:06:03.240 from a political standpoint. The complicated nature of obscenity laws 0:06:03.279 --> 0:06:05.880 in the United States meant that it would be very 0:06:05.880 --> 0:06:10.080 difficult to make sure that Netflix was complying with laws 0:06:10.160 --> 0:06:13.480 on a state by state basis, because what is permitted 0:06:13.520 --> 0:06:16.119 in one state might be against the law in another state. 0:06:16.160 --> 0:06:19.360 And when you are a company that is trying to 0:06:19.440 --> 0:06:24.120 have a uh, you know, a country wide presence, a 0:06:24.240 --> 0:06:28.159 nationwide presence, it just didn't make any sense to carry 0:06:28.200 --> 0:06:32.120 that kind of content. On the marketing side, Netflix reached 0:06:32.120 --> 0:06:36.159 out to early Internet influencers like Harry Knowles, who was 0:06:36.200 --> 0:06:38.880 the founder and still is the founder of Ain't Cool 0:06:39.000 --> 0:06:42.159 dot com or Ain't It Cool News, And in those days, 0:06:42.200 --> 0:06:46.040 Ain't It Cool was growing beyond being a niche film 0:06:46.160 --> 0:06:48.880 fans site where you know, it kind of started out 0:06:48.920 --> 0:06:50.120 as a bunch of a place where a bunch of 0:06:50.160 --> 0:06:53.680 film geeks would get together and argue about movies and 0:06:53.839 --> 0:06:58.240 read various articles about upcoming films and rumors and things 0:06:58.320 --> 0:07:01.800 like that. And at this point it was growing beyond that. 0:07:01.839 --> 0:07:05.040 It was getting attention from directors and movie studios because 0:07:05.040 --> 0:07:09.520 the industry, the film industry had seen that Knowles and 0:07:09.640 --> 0:07:12.840 his staff that we're writing for and it Cool had 0:07:12.960 --> 0:07:17.240 enormous reach, and so for a relatively modest investment on 0:07:17.280 --> 0:07:20.640 the part of the studio, those companies could get a 0:07:20.680 --> 0:07:22.840 lot of marketing benefits. It was like a lot of 0:07:22.920 --> 0:07:28.000 word of mouth marketing through this website. So just by 0:07:28.200 --> 0:07:30.160 doing a little bit of investment in the website they 0:07:30.160 --> 0:07:33.400 could get a big payoff. So Netflix followed suit and 0:07:33.480 --> 0:07:36.400 invested money in and at cool dot com, and in return, 0:07:36.840 --> 0:07:40.200 Noels and his crew would frequently mention the service. Another 0:07:40.240 --> 0:07:43.720 clever move that Netflix made early on was they negotiated 0:07:43.760 --> 0:07:48.080 with DVD manufacturers to include a coupon for a free 0:07:48.120 --> 0:07:52.400 Netflix rental with new DVD players. So if you went 0:07:52.440 --> 0:07:54.840 out you bought a DVD player back in those days, 0:07:55.040 --> 0:07:57.680 inside the box would be a little coupon that would 0:07:57.680 --> 0:08:01.680 allow you to get a free rental off Netflix. So 0:08:01.880 --> 0:08:05.960 these promotions were drawing attention, but they were also costing 0:08:06.080 --> 0:08:10.320 Netflix money. The company was losing cash on every transaction. 0:08:10.680 --> 0:08:13.920 For one thing, the labor costs were very high, and 0:08:13.960 --> 0:08:17.600 the promotions were also cutting into revenue. They also had 0:08:17.680 --> 0:08:24.520 made a pretty terrible rookie mistake, so they didn't spend 0:08:24.560 --> 0:08:27.440 the money to build out a way to handle unique 0:08:27.520 --> 0:08:30.880 coupon codes. They had no system in place to create 0:08:30.960 --> 0:08:35.400 a sequential coupon code so that each code was unique 0:08:35.400 --> 0:08:37.880 and could only be used once, so they shipped out 0:08:37.920 --> 0:08:42.000 coupons that did not have unique identifiers on them. That 0:08:42.040 --> 0:08:46.840 meant that if you were unscrupulous, you could duplicate those 0:08:46.880 --> 0:08:50.400 coupons and use coupons multiple times, because you could just 0:08:50.480 --> 0:08:54.679 keep making copies and said each coupon appeared to be legitimate. 0:08:54.760 --> 0:08:57.760 Netflix would ship out rentals each time they got a coupon, 0:08:57.960 --> 0:09:02.640 so people were essentially scaming Netflix, and Netflix was losing 0:09:02.640 --> 0:09:06.600 money in the process. In August, Randolph was able to 0:09:06.640 --> 0:09:09.960 secure six million dollars in venture capital to keep the 0:09:10.040 --> 0:09:15.319 company afloat, but that would only managed to save Netflix 0:09:15.559 --> 0:09:17.800 until the end of ninety eight. Something else was going 0:09:17.840 --> 0:09:21.160 to have to happen. The one part in Netflix's business 0:09:21.280 --> 0:09:26.319 that was doing fairly well was in DVD sales, not rentals. 0:09:26.760 --> 0:09:29.959 There weren't enough customers who would return to rent a 0:09:30.080 --> 0:09:33.720 DVD after they had used up their free rentals they 0:09:33.960 --> 0:09:36.720 go through, They'd rent a couple of top titles for free, 0:09:37.160 --> 0:09:41.840 and then they wouldn't use the service anymore. So that 0:09:42.040 --> 0:09:46.680 was not really a revenue starter, but the sales were 0:09:46.679 --> 0:09:50.440 working better. However, that was a short term win, and 0:09:50.559 --> 0:09:52.840 Randolph knew it. He knew that it was not going 0:09:52.880 --> 0:09:56.800 to be long until you saw other big companies, established 0:09:56.840 --> 0:10:00.000 companies things like Amazon dot Com get into that space 0:10:00.000 --> 0:10:04.920 ace and start selling DVDs, and the retail business would 0:10:05.000 --> 0:10:08.240 be very crowded and Netflix, being a small time player, 0:10:08.280 --> 0:10:12.600 would be unable to uh to compete in that space, 0:10:13.040 --> 0:10:15.080 so he knew that that was not a way to 0:10:15.280 --> 0:10:20.760 build the future of a business. In early Read, Hastings 0:10:20.800 --> 0:10:24.360 began to become more involved in Netflix. Remember, up to 0:10:24.400 --> 0:10:27.920 that point, he had largely been hands off. He had 0:10:27.960 --> 0:10:31.240 provided the initial seed money for the company, and he 0:10:31.280 --> 0:10:35.160 had brainstormed with Mark Randolph when they were first talking 0:10:35.160 --> 0:10:37.840 about what Netflix would be and how it would be 0:10:37.880 --> 0:10:40.920 the Amazon dot Com of video rental, but apart from that, 0:10:40.960 --> 0:10:45.640 he really hadn't beencome very involved. Not long after Read, 0:10:45.679 --> 0:10:48.560 Hastings decided he was going to get his hands dirty 0:10:48.559 --> 0:10:51.160 and Netflix. He announced that he and Randolph would become 0:10:51.320 --> 0:10:55.160 co chief executives of the company, a move that apparently 0:10:55.240 --> 0:10:59.400 didn't please Mark Randolph very much. Hastings immediately set the 0:10:59.440 --> 0:11:02.760 tone for his style at this meeting, as he then 0:11:02.920 --> 0:11:08.280 fired the brand new HR manager, Lisa Battalia Rice. He 0:11:08.400 --> 0:11:10.760 announced that he was going to bring in his own 0:11:11.000 --> 0:11:14.920 HR manager, Patty McCord. And McCord had previously worked with 0:11:14.960 --> 0:11:19.560 Hastings at Pure Atria to run the human resources department there, 0:11:19.960 --> 0:11:23.040 according to the book Netflix, which was a great resource. 0:11:23.520 --> 0:11:26.800 This may have been because McCord was seen as someone 0:11:26.840 --> 0:11:29.800 who could act as kind of a liaison between Hastings 0:11:29.800 --> 0:11:33.959 and staff, because Hastings took a very engineering approach to 0:11:34.120 --> 0:11:37.840 problems and was very blunt and could be quite critical, 0:11:38.280 --> 0:11:41.400 so she would sort of soften the edges a little bit. 0:11:42.200 --> 0:11:46.640 According to again people within the company, Randolph in Hastings 0:11:47.000 --> 0:11:51.400 were very different kinds of leaders. Mark Randolph was more 0:11:51.679 --> 0:11:56.760 empathetic and more outgoing. Read Hastings was more concerned with 0:11:56.800 --> 0:12:01.439 efficiency and performance, and if you weren't performing to expectations, 0:12:01.760 --> 0:12:06.480 you were alerted to that fact pretty quickly. Hastings was 0:12:06.760 --> 0:12:11.560 more quantitative, Randolph more qualitative. Hastings was said to have 0:12:11.640 --> 0:12:15.280 an engineering approach to pretty much any problem, including people problems, 0:12:15.440 --> 0:12:18.640 whereas Randolph was more about allowing people to explore creative 0:12:18.640 --> 0:12:23.360 approaches to challenges. And Hastings was a big hit over 0:12:23.440 --> 0:12:26.959 on the software engineering side of the business. Mark Randolf 0:12:27.000 --> 0:12:29.680 was more of a hero on the marketing and public 0:12:29.720 --> 0:12:33.800 relations and business plans side. By the end of the year, 0:12:34.440 --> 0:12:38.880 Read Hastings would assume the position of Soul CEO and 0:12:38.960 --> 0:12:42.480 bomp Mark Randolf down to president. Then a little later, 0:12:42.520 --> 0:12:45.440 Read Hastings would also take on the role of president 0:12:45.840 --> 0:12:49.720 and Mark Randolph became an executive producer. Hastings was also 0:12:49.800 --> 0:12:53.240 the chairman of the board. Mark Randolph had a position 0:12:53.400 --> 0:12:56.600 on the board that Read Hastings originally wanted to free 0:12:56.720 --> 0:13:01.080 up to UH to grant to an investor. Mark Randolph refused, 0:13:01.120 --> 0:13:05.720 said he absolutely insisted on maintaining his seat at the board, 0:13:05.800 --> 0:13:10.080 and Read Hastings relented at that point. Now, in part, 0:13:11.600 --> 0:13:17.520 Randolph kind of accepted this fate because he knew that 0:13:17.600 --> 0:13:20.800 Read Hastings was important. He knew that he needed Hastings 0:13:20.840 --> 0:13:24.640 in order to keep Netflix afloat. Because investors knew who 0:13:24.679 --> 0:13:29.559 Hastings was. Hastings had founded Pure Software. Pure Software had 0:13:29.600 --> 0:13:32.160 merged with a company called Atria, and the company Pure 0:13:32.240 --> 0:13:36.400 Atria was a huge success for investors. It got acquired 0:13:36.920 --> 0:13:40.640 for a massive deal that netted investors a huge amount 0:13:40.640 --> 0:13:43.320 of money. So Hastings was really popular with these guys. 0:13:43.880 --> 0:13:46.960 They said, this guy has made me money before, and 0:13:47.000 --> 0:13:48.920 if he says this company is worth investing in, I'm 0:13:48.920 --> 0:13:51.520 going to pay attention. So I gave Netflix a better 0:13:51.640 --> 0:13:54.640 chance at getting investment money if Hastings was the one 0:13:54.720 --> 0:13:58.520 in the spotlight. Hastings made a couple of controversial decisions 0:13:58.640 --> 0:14:02.000 in those early months, he declared that Netflix would stop 0:14:02.000 --> 0:14:04.600 selling DVDs and would have to find a way to 0:14:04.679 --> 0:14:08.880 make renting DVDs online profitable or the company would just 0:14:08.920 --> 0:14:11.320 have to close up shop. Then he reached out to 0:14:11.440 --> 0:14:15.679 Amazon dot Com and he suggested that Netflix direct customers 0:14:15.720 --> 0:14:18.640 to Amazon if they were interested in buying a film, 0:14:18.960 --> 0:14:22.800 and in return, Amazon dot Com would display ads for Netflix. 0:14:23.440 --> 0:14:25.720 So this was a deal that was struck up, and 0:14:25.800 --> 0:14:28.400 some of the folks over at Netflix, over in marketing, 0:14:28.560 --> 0:14:32.200 thought this was a colossal mistake because they thought that 0:14:32.280 --> 0:14:37.040 Netflix was effectively handing customers over to a competitor. Netflix 0:14:37.080 --> 0:14:40.840 also had early meetings with Hollywood Video and Blockbuster, the 0:14:40.920 --> 0:14:45.320 two big names in video rental VHS rental, and they 0:14:45.360 --> 0:14:47.840 were kind of feeling things out to see if either 0:14:47.880 --> 0:14:52.000 company might be interested in purchasing Netflix. Because again, one 0:14:52.000 --> 0:14:54.200 way that startups make a lot of money is not 0:14:54.360 --> 0:14:57.560 through creating a sound business plan, but by getting acquired 0:14:57.560 --> 0:15:02.840 by a larger company. Uh. Now, if those companies were 0:15:02.880 --> 0:15:06.880 at all interested in acquiring Netflix, it certainly wasn't at 0:15:06.920 --> 0:15:12.280 the price that Hastings and Randolph were willing to entertain. Uh. 0:15:12.320 --> 0:15:14.600 They were asking for a price much higher than what 0:15:14.760 --> 0:15:19.160 either company would be willing to to meet, and neither 0:15:19.240 --> 0:15:23.360 companies seemed particularly interested in any sort of collaborative efforts. 0:15:23.400 --> 0:15:26.680 Both companies turned down Netflix's offer. At the end of 0:15:27.680 --> 0:15:31.720 Netflix posted a loss of eleven million dollars and there 0:15:31.720 --> 0:15:35.320 were tons of free redual coupons still out there to 0:15:35.600 --> 0:15:38.920 do more damage to the company. So something was going 0:15:38.960 --> 0:15:41.920 to have to change. What was it? I'll tell you 0:15:41.960 --> 0:15:44.280 in a minute, but first let's take a quick break 0:15:44.520 --> 0:15:55.000 to thank our sponsor Read Hastings was able to keep 0:15:55.080 --> 0:15:59.520 Netflix going largely by getting investments from venture capitalists. Over 0:15:59.840 --> 0:16:02.320 the year and a half period, he would raise more 0:16:02.360 --> 0:16:06.800 than one hundred million dollars in investments. At the same time, 0:16:07.400 --> 0:16:11.960 he was changing the corporate culture. They company had started 0:16:12.000 --> 0:16:15.640 as sort of this chaotic creative environment with all these 0:16:15.680 --> 0:16:18.000 different people trying to figure out, how can we make 0:16:18.000 --> 0:16:20.720 a business plan that makes sense, how can we market it, 0:16:20.800 --> 0:16:24.040 how can we attract customers, how can we actually create 0:16:24.160 --> 0:16:27.840 the infrastructure that will allow it to happen. And it 0:16:28.000 --> 0:16:32.320 was uh sort of a crazy but really energetic place 0:16:32.480 --> 0:16:36.600 to work, and Hastings was changing things and buttoning things 0:16:36.640 --> 0:16:39.760 down and making it more hierarchical. Many of the founding 0:16:39.760 --> 0:16:44.880 team members for Netflix left the company in including Jim Cook, 0:16:45.000 --> 0:16:49.920 the the finance and operations director, Christina Kish left, Terse 0:16:50.040 --> 0:16:53.920 Smith left, and more. Joining the company was a guy 0:16:54.000 --> 0:16:58.080 named Tom Dylan, who took over operations from the outgoing 0:16:58.160 --> 0:17:01.920 Jim Cook. Dylan had actually worked at a company called 0:17:02.080 --> 0:17:06.280 c Gate, and he headed up an international operations division 0:17:07.240 --> 0:17:13.080 magnitudes larger than Netflix, huge h division compared to what 0:17:13.160 --> 0:17:17.160 Netflix was doing. Uh. Netflix presented a new challenge at 0:17:17.200 --> 0:17:21.399 a tiny scale in comparison, but Dylan needed to do 0:17:21.440 --> 0:17:24.400 some dramatic changes to really turn things around for the company. 0:17:24.520 --> 0:17:28.320 When he came on board, it cost Netflix six dollars 0:17:28.400 --> 0:17:31.679 to as symbol and ship and order. Now, most orders 0:17:31.960 --> 0:17:35.359 would actually include three discs at a time, so it 0:17:35.440 --> 0:17:38.800 wasn't just six dollars per disc. It was six dollars 0:17:38.800 --> 0:17:42.879 per order, but still was not an efficient system if 0:17:42.920 --> 0:17:46.760 you think like four dollars per disk, and so that's 0:17:46.760 --> 0:17:49.840 like twelve dollars per order and it costs you six 0:17:49.880 --> 0:17:52.440 dollars to send the order out. That's not great. So 0:17:52.640 --> 0:17:54.760 Dylan wanted to get the order cost down to a 0:17:54.880 --> 0:17:58.280 third of what it was. One thing he did was 0:17:58.320 --> 0:18:01.200 to change the shipping process so that each envelope would 0:18:01.200 --> 0:18:03.760 hold only one disc at a time. And that seems 0:18:03.760 --> 0:18:06.960 counterintuitive if you think, wait, if it costs six dollars 0:18:06.960 --> 0:18:10.880 to fill an envelope, but this actually cut down on costs. 0:18:10.960 --> 0:18:14.520 For one, it would let Netflix send out disks when 0:18:14.560 --> 0:18:17.400 they were available, instead of holding an order for all 0:18:17.520 --> 0:18:19.960 three discs to be ready at the same time. So 0:18:20.320 --> 0:18:23.360 if a customer is waiting on three different titles, then 0:18:23.440 --> 0:18:26.200 you can at least send one title out ahead while 0:18:26.240 --> 0:18:29.560 the other two are either coming back to Netflix or 0:18:29.600 --> 0:18:33.080 haven't yet been processed, and you can get orders in 0:18:33.119 --> 0:18:36.520 people's hands faster. Secondly, it brought the cost down on 0:18:36.600 --> 0:18:40.480 individual shipping. The individual envelopes were cheaper to send than 0:18:40.520 --> 0:18:43.439 the larger ones that would hold three disks, and he 0:18:43.480 --> 0:18:46.280 included a barcode system to help sort orders in the 0:18:46.320 --> 0:18:51.320 warehouse and increase efficiency. The next bit is particularly interesting 0:18:51.359 --> 0:18:54.800 to me. Dylan and his team discovered that people living 0:18:54.880 --> 0:18:58.920 within a one day delivery zone were far more likely 0:18:58.960 --> 0:19:02.000 to recommend the serve as to friends. So if you 0:19:02.119 --> 0:19:05.920 live someplace where you would get your order a day 0:19:05.960 --> 0:19:10.120 after you ordered it, you were more likely to be 0:19:10.760 --> 0:19:14.199 a loyal customer and to bring other people into the fold. 0:19:14.800 --> 0:19:17.720 So this helped with customer attraction and retention. And so 0:19:18.280 --> 0:19:21.800 Dylan got hold of a c D that held information 0:19:22.800 --> 0:19:26.160 about US Post Office delivery data for ZIP codes across 0:19:26.200 --> 0:19:29.280 the United States, and he headed up the creation of 0:19:29.320 --> 0:19:34.080 software that could identify the ideal locations to establish a 0:19:34.119 --> 0:19:38.480 distribution center in various regions across the US. And that way, 0:19:38.680 --> 0:19:41.360 he would say, all right, well, we see that there's 0:19:41.400 --> 0:19:45.639 a very dense population in let's say New York City. 0:19:45.880 --> 0:19:49.840 Where can we put a distribution center where we can 0:19:49.880 --> 0:19:53.119 guarantee one day delivery to as many people in that 0:19:53.200 --> 0:19:56.360 region as possible. And so they started taking all these 0:19:56.440 --> 0:19:58.560 data points and they started to look at maps, and 0:19:58.560 --> 0:20:01.400 they started to actually plan out the physical locations where 0:20:01.400 --> 0:20:04.760 distribution centers should be to be able to get one 0:20:04.840 --> 0:20:07.720 day delivery to the largest number of potential customers for 0:20:07.800 --> 0:20:11.400 any given region. Dylan also figured out that in order 0:20:11.440 --> 0:20:15.240 to support a distribution center, each center would need about 0:20:15.320 --> 0:20:21.320 fifteen thousand customers two to justify the existence of that center. 0:20:21.440 --> 0:20:25.119 So using these different points of data and this data 0:20:25.160 --> 0:20:28.240 analysis approach, they were able to point out a map 0:20:28.280 --> 0:20:30.960 and say here's where we need to build our next 0:20:31.200 --> 0:20:34.399 distribution center, and then they could guarantee one day delivery 0:20:34.440 --> 0:20:36.760 to customers in that region and build up the base 0:20:36.800 --> 0:20:39.679 of users they would need to generate regular revenue. And 0:20:39.720 --> 0:20:42.720 it's that kind of data analysis that I find really interesting. 0:20:42.720 --> 0:20:45.639 It's a very innovative approach to figuring out where to 0:20:45.760 --> 0:20:49.480 put your operations as a company. In the summer of 0:20:51.160 --> 0:20:54.480 Netflix began to experiment with a couple of different business 0:20:54.520 --> 0:20:58.720 models to get away from the per title rental approach 0:20:58.760 --> 0:21:02.359 the company had been using up to that point. So remember, 0:21:02.440 --> 0:21:04.159 up to this point, if you wanted to run a 0:21:04.160 --> 0:21:07.600 moview in Netflix, you're paying on a per title basis. 0:21:07.600 --> 0:21:09.320 It might be four dollars per title. It might be 0:21:09.359 --> 0:21:11.840 five dollars per title if you're running three movies, and 0:21:11.920 --> 0:21:15.239 that's twelve bucks or fifteen bucks, depending on you know, 0:21:15.440 --> 0:21:19.399 how how expensive each title was, and they were looking 0:21:19.400 --> 0:21:23.040 at alternatives to this approach. One of these was called 0:21:23.080 --> 0:21:26.480 the home Rental Library, So for twenty dollars per month, 0:21:26.880 --> 0:21:29.520 customers would be able to rent up to six movies 0:21:29.600 --> 0:21:31.520 at a time and keep them for as long as 0:21:31.600 --> 0:21:34.200 they wanted. There'd be no late fees, so if you 0:21:34.280 --> 0:21:36.840 kept them for three months, that's fine, and when you 0:21:36.880 --> 0:21:39.920 return them, and if you're paying the membership fee month 0:21:39.960 --> 0:21:43.240 to months, then you could get six more movies. The 0:21:43.280 --> 0:21:47.399 other approach was called serialized delivery. Now, in that version, 0:21:47.520 --> 0:21:52.159 a customer would first receive whatever the next available film 0:21:52.240 --> 0:21:54.960 is in his or her Q. I talked about the 0:21:55.000 --> 0:21:58.440 queue in the previous episode. That's the list of movies 0:21:58.480 --> 0:22:01.760 you've expressed interest in. And the more interested you are 0:22:01.800 --> 0:22:03.480 in a movie, the higher up you should put it 0:22:03.560 --> 0:22:06.600 in your queue. And so what would happen is you 0:22:06.680 --> 0:22:09.359 get a film, you'd watch your movie, you return it 0:22:09.400 --> 0:22:13.760 to Netflix. Netflix would scan the returning disk, they would 0:22:13.800 --> 0:22:17.000 register that you had returned your movie. They would look 0:22:17.040 --> 0:22:20.280 at your queue, and whatever the next available title in 0:22:20.359 --> 0:22:23.440 theory was that was on your queue, they would send 0:22:23.480 --> 0:22:28.000 that one to you next and uh. Then these both 0:22:28.040 --> 0:22:33.480 were really popular. Hastings ordered actually, that both serialized delivery 0:22:33.520 --> 0:22:36.359 and the home rental library and the new queue feature 0:22:36.600 --> 0:22:39.600 should all be tested at the same time, which drove 0:22:39.640 --> 0:22:42.920 marketing nuts because they wanted to be able to isolate 0:22:43.480 --> 0:22:47.359 these different approaches and to measure them against each other. 0:22:47.960 --> 0:22:51.000 It turned out the customers loved all three features, so 0:22:51.119 --> 0:22:55.080 Randolph decided that they should combine them into a single offering. 0:22:55.960 --> 0:22:59.679 Customers would pay a monthly subscription fee. This was an 0:22:59.680 --> 0:23:03.760 alter her native to the per title rental fee. It 0:23:03.960 --> 0:23:08.520 was not mandated originally, it was an alternative. These subscriptions 0:23:08.560 --> 0:23:11.880 would be in tears different levels that would allow customers 0:23:11.880 --> 0:23:15.320 to choose how many DVDs they could rent at a time, 0:23:15.359 --> 0:23:18.520 and each tier would cost a different amount. You know, 0:23:18.560 --> 0:23:20.200 the more movies you wanted to be able to check 0:23:20.240 --> 0:23:22.480 out at once, the higher you would pay per month. 0:23:23.440 --> 0:23:26.200 Then there would be no late fees, and upon returning titles, 0:23:26.240 --> 0:23:30.040 you would get the next available DVD from your queue 0:23:30.600 --> 0:23:32.560 that would be the next one to be sent along again. 0:23:32.600 --> 0:23:37.120 In theory, it's probably from this point that Hastings decided 0:23:37.160 --> 0:23:40.720 to go with the version of Netflix's history that was 0:23:40.720 --> 0:23:44.560 passed around as gospel for several years, and that was 0:23:44.680 --> 0:23:46.880 the version that I talked about in the last episode 0:23:46.880 --> 0:23:49.560 where he had said he got the idea for Netflix 0:23:49.600 --> 0:23:53.560 after he got a late fee for keeping Apollo thirteen 0:23:53.720 --> 0:23:57.600 from a blockbuster for too long. In reality, Netflix didn't 0:23:57.680 --> 0:24:00.119 go with this no late fee approach until it was 0:24:00.119 --> 0:24:03.199 about a year and a half old, so that was 0:24:04.000 --> 0:24:08.359 not a a key part of Netflix from day one. 0:24:08.480 --> 0:24:12.840 It was well after they had begun to do business 0:24:12.880 --> 0:24:17.480 that they adopted that philosophy. The new model that they 0:24:17.560 --> 0:24:22.359 used would debut on September. The plan would allow a 0:24:22.359 --> 0:24:25.760 customer to receive up to four titles simultaneously for the 0:24:25.840 --> 0:24:30.360 monthly subscription rate of fifteen dollars and nine cents, which 0:24:30.400 --> 0:24:34.480 I think is too low to qualify as a princely sum, 0:24:34.520 --> 0:24:38.760 but the rollout was targeted. Most visitors would not see 0:24:38.880 --> 0:24:41.800 this offer at all. They would just see the normal 0:24:41.920 --> 0:24:45.760 per title rental structure. But a few people did see 0:24:45.760 --> 0:24:48.719 the offer, and the people who saw it tended to 0:24:48.800 --> 0:24:51.639