Netflix's product expertise is at the heart of everything it does, including why and how it makes its original content. This culture is easily overlooked, but doing so can be fatal. Netflix is as much a tech and product company as Facebook, Google or Amazon.

There’s a pernicious and persistent narrative about Netflix where the company’s success is overwhelmingly attributed to the mistakes of its suppliers. Not only did these suppliers (a group that included nearly every major media company) continually sell the most valuable rights to their most valuable content to Netflix, they massively underpriced these deals. As such, the streaming upstart was able to (1) access large volumes of high quality content at a time when it had none of its own; (2) build a business atop the creative successes of its eventual competitors; and (3) benefit from years of relatively uncontested OTT leadership. Hence success!

These deals absolutely enabled Netflix to become the comptwany it is today. Without them, Netflix would have had nothing to offer customers, nothing to learn from its customers’ behaviors and no history with which to attract talent needed to produce Netflix Originals. But this narrative is offensively incomplete. Netflix’s success required three other things, each of which is critical to understanding the company’s trajectory, especially as more suppliers attempt to withhold content.

Creative success, for one, shouldn’t be overlooked. Netflix’s first and third Originals (House of Cards and Orange Is the New Black) were huge hits with critics, audiences and the major award shows. And while Netflix now produces more Originals than any other network, its hit rate remains strong. To that end, Netflix’s volume (~10% of all Original scripted series that aired in United States last year, depending on your definition) doesn’t explain having three of seven Best Drama nominations and two of seven Best Comedy nominees at last year’s Emmys. In addition, Reed Hastings’ aggressive investment [see Part 1 of this series] in long-term potential, rather than short-term profits or Wall Street appeasement, is inextricable from the company’s present-day success.

However, the third and most commonly overlooked explanation for Netflix’s success is its most differentiated and defensible one: product. At a time in which most tech companies need to be bullied into admitting they’re also media companies, Netflix’s tech identity is often glossed over. It is as much a technology and product company as Google, Apple or Amazon.

Netflix has built a vast set of technologies for its business and each is arguably best-in-class. This includes proprietary video compression algorithms (which optimize not just for the bandwidth available, but for context specific needs of the content – My Little Pony is not the same as The Avengers, nor is an action scene in The Avengers the same as a dialogue-centric one); distribution (Netflix’s apps are available fully featured on more devices worldwide than any other video providers, even devices such as the Nintendo 3DS and select treadmills); A/B tested covert art that leverages every frame of a show (with the frames pre-filtered for suitability via computer vision); a suite of engagement functions such as autoplay trailers (themselves algorithmically generated for licensed titles), auto-resume when selecting a previously started title, auto-play next episode, credits skipping and vertical video trailers; personalization (Netflix was the first OTT video service to rollout user accounts) and recommendations that tap into 80,000 tagged genre sub-categories; interactive/dynamic storytelling technology; and even home-grown SaaS applications designed to improve the production of Netflix Originals by third parties. The list goes on.

Netflix’s product investments stem from its obsession with engagement. This pursuit is most simply distilled into the metric “hours per subscriber per month,” which Hastings often identifies as the KPI most strongly associated with customer retention and pricing. To this end, each of the above technologies lubricates access to, discovery on, or consumption of Netflix. These benefits are generally valuable, but for Netflix, they were (and remain) essential. Not only was the company’s catalog primarily composed of rerun content, but it saw a limited runway in which to develop its own library of premium original TV – and had to do so during a time with an unprecedented competition for consumer attention and against an ingrained Pay TV ecosystem with 90% penetration and an average of 8+ hours of use per household per day. Last year, for example, consumers could choose from not just 500 original scripted series airing new seasons (up from 150 in 1999), they could also access the entire accumulated history of television. In this marketplace, there’s a significant competitive advantage in being the best and easiest place to watch video (the consumer view), and the best place to get a video watched (the supplier one). Few would disagree that Netflix has achieved both.

Similarly, Netflix’s scale and technology enables it to launch a mediocre show to a larger audience than most of its competitors can when they produce an outstanding one. This doesn’t mean the company doesn’t want great shows, but it does mean that the risk of producing a disappointing one is much lower, and the competitive risk of losing out on a promising script is moderated.

These two advantages also produce a virtual cycle of success (or arbitrage value, if you prefer) that enables Netflix to economically outspend its competitors when bidding on content. As such, the service is well positioned to scoop up much of the content held by suppliers who get cold feet on their own direct-to-consumer offerings (with this strategy being at the core of Netflix bear hypotheses). In some instances, Netflix’s pricing power will even induce this hesitation. In March, for example, The Information reported that after launching SeeSo (a comedy-centric SVOD service), NBCUniversal considered pulling The Office off Netflix to make it exclusive to its new service. In response, Netflix offered to raise its license fee by tens of millions of dollars – but only for exclusivity to its platform. NBCUniversal accepted the deal and a year later shut down SeeSo. While much of Netflix’s tech can and will be replicated, the momentum and scale advantages borne from its advances will endure. And Netflix will continue to make technological and experiential gains as its competitors replicate their older achievements.


Content as Product

Netflix’s obsession with engagement is quintessentially tech. It’s far more analogous to Facebook’s north star than, say, that of FX or HBO. To point, Netflix’s product mentality is most clear when it’s in opposition to the sanctity or rights of a creator. Algorithmically generated trailers and cover art, for example, wrest control from creatives and hand it over to a computer processor. When combined, opening-titles skipping and auto-play next episode deprive talent of all on-screen recognition (this is a serious issue: In 1977, George Lucas was fined the present-day equivalent of $1MM and forced to resign from the Directors’ Guild of America after putting Star Wars credits at the end of the film, rather than in the opening titles and after the film). Binge releases and auto-play, meanwhile, treat every episode as a delivery mechanism for the next.

We also see this in the content itself. Netflix’s original series are continually criticized for being bloated or too long – six-episode stories are routinely stretched to 10-13 episodes to elongate view time (even if it drags down average quality). Similarly, Netflix has taken flak for the (alleged) quality of its high-budget Original films, such as the December 2017 release Bright (27% on Rotten Tomatoes), January 2018’s The Cloverfield Paradox (16%), February’s Mute (13%) and March’s The Titan (18%) and The Outsider (20%) – a claim the company rebuffs by brandishing its (allegedly) massive viewership numbers.

This is not to say Netflix doesn’t value quality – again, it makes much of the best TV content available. And all things being equal, a great show is more valuable to Netflix than a good one. But the “all things being equal” part is critical. When engagement (i.e. hours) is the top priority, incentives change. Quality is only one of many important variables – and probably not the top one. Is it better for 125 people to watch a six-episode season of Jessica Jones that’s of ‘A-’ quality, for example, or for 100 to watch a 12-episode version that’s dragged down to a ‘B’? After seven Marvel series, Netflix’s answer seems to be the latter.


Netflix’s Engagement Race

The prioritization of engagement time over quality is controversial, but there are a few explanations. To start, one has to assume Netflix is correct in observing that, at least in the short-run, watch time has a (much) stronger impact on retention than quality (and of course, the former is a more objective, quantifiable and analyzable metric). This relationship likely stems from the unique dynamics of an unbundled, D2C subscription content service. The consumer mentality of going to a movie theater and buying a ticket (or even watching a live airing of a show that airs precisely from 9–10 PM and is 30% advertising) has a different quality threshold than a video that’s part of an All You Can Eat offering. As a recurring subscription, Netflix is more about value than optimizing for satisfaction per minute. To the consumer, renewal is a question of “Did I get enough satisfaction for $10?”; for Netflix, it’s “Do we provide enough value to charge another dollar?” In addition, each hour watched provides an additional opportunity for Netflix to promote other content on its service (good for engagement) and cannibalize a competitor’s watch time (many see OTT video as “winner takes most”).

What’s more, the Jessica Jones example above is likely exaggerated. Few viewers will abandon a show after six episodes in a binge-released, on-demand environment. Even if it’s not great, they’ll typically stick with it (thanks, autoplay!). To this end, the right case is more likely 125 people watching six hours versus 118 watching 12. It’s also worth mentioning that Netflix’s “job to be done” isn’t necessarily to provide the best possible content. Americans watch on average 5.5 hours of video each day, and it isn’t all intended to be mentally challenging or creatively inspiring. Some of it simply helps one relax or pass the time (“treadmill TV” as REDEF founder Jason Hirschhorn puts it). Not all viewers want “the best possible Jessica Jones experience” – some just like spending time in its universe. Accordingly, many would rather enjoy 13 hours of the series than seven, even if the quality dips. (And Netflix does have a large catalog of creative, stimulating content they can switch to at anytime)

You can see the idea of content-as-engagement play out when Netflix discusses content investments. On its “Top Investor Questions” page, Netflix answers the question “How do you evaluate new content deals or renewals?” this way:

We utilize detailed statistical models to determine expected hours of viewing for each piece of content over its license period. We compare cost per hour viewed against other “like” content deals (i.e. exclusive versus non-exclusive, TV versus movies, etc.) We look for high engagement and cost efficiency. For renewals, we look to renew content that performs well (based on hours generated relative to the cost) and do not renew content where the price doesn’t make sense relative to the value generated. We feel we have good breadth of content so that no specific title or set of titles is must-renew.

This view considers content as fundamentally substitutable – because it’s not an experience being bought (or sold), it’s time. Quality is expressed through viewing volume and, as with most substitutable goods, pricing efficiency is paramount. If the average title generates 100 hours per dollar, then a title that generates only 80 hours costs Netflix 25% of potential viewing hours and thus avoidable subscriber losses and realizable subscriber gains. This dynamic is further bolstered by the role of cost amortization. The decision to make The Crown is an expensive one irrespective of the number of hours produced; set building, costume design, casting, scoring and location scouting are upfront, fixed costs, largely independent of episode count. As such, a 10-episode season typically won’t cost 11% more than a nine-episode one. Given the likelihood that a viewer would watch ten episodes rather than nine if given the choice, elongation drives both net engagement and efficiency gains. And that’s just in adding one episode.


The Engagement Marathon

There’s a way to read the above that makes Netflix sound creatively indifferent. However, the company’s incentives are not altogether dissimilar from those of the traditional ecosystem. On broadcast and basic cable TV, each additional minute of watch time generates direct revenue (via advertising), for example. In an ad-free, subscription environment, this isn’t the case – but if view time drives retention and pricing power, the consequence isn’t that different. It just manifests a little differently.

In the linear world, most networks target seasons of 20-24 episodes – much longer than even Netflix’s longest originals. This length was in part because the average viewer only watched a third of a linear season. After all, few fans could tune in every Wednesday at 9PM (especially after accounting for sporadic weeks of reruns and pre-emption by events like the World Series or State of the Union). As such, networks needed lengthy seasons to properly engage/retain a show’s fans and to monetize their most successful properties. In addition, series syndication deals were largely paid on a per episode basis, with extra episodes therefore justified by clear and reliable business cases. However, traditional networks faced not just different length motivations, but also different creative dynamics. Each episode had to stand alone, sustain attention through multiple commercials in a competitive timeslot (at a time when the choice to watch X meant missing Y), and convince viewers to come back exactly 167 hours later for more. As a result, seasons might have been long, but narratives remained tight and a few droning episodes could imperil a series. The advent of time-shifted consumption (DVR, TVE, SVOD) has enabled more complex stories (and a broader shift from episodic to serialized TV), but it has also lessened the cost of padding a season or storyline. You might not come back for a weekly show after three mediocre episodes, but you might keep plowing through a slow one on a Sunday afternoon when it’s available on-demand, binge-released and ad-free. In total, then, Netflix’s rationale and use of over-long seasons may be unique, yet still consistent with both the industry it preceded and the one that it has remodeled.

Crucially, however, Netflix’s streaming incentives are distinct from those affecting its primary digital-era competitors. Adding value to Apple Music, Amazon Prime or a multi-dimensional Disney product is a rather different business model than driving a standalone video service. Apple has expressed no interest in licensed content (which is effective at driving bulk viewer hours) or competing on volume. Instead, the company is focusing on releasing eight to 10 high-budget, high-profile series a year. Amazon’s strategy has historically focused on maximizing impact per viewer, rather than raw engagement or total viewership. Disney’s forthcoming streaming service will have a decent-sized library, but again, the effort is focused on having the most desirable content in the world.

Similarly, HBO, Showtime and Starz are more focused on having one must-see show on at all times, with a sprinkling of must-see movies or events throughout the year. FX’s business model is centered around its differentiated content style and awards, both of which make it an essential part of any TV bundle (which drives both distribution and pricing power). These four players, each of which is seen as Netflix’s primary “traditional TV” competitors, pursue weekly (not binge) releases and produce only a dozen original series each per year. Growing watch time through additional volume might actually erode profit margins, if not total profit (else we’d have already seen programming expansions). Still, it’s not clear how traditional TV’s major brands will adapt to an unbundled, direct-to-consumer marketplace (each of these networks benefits from both Pay TV bundling and/or bundling with one another). To defeat binge-and-churn subscriptions and grow their subscriber bases (HBO, Showtime and Starz have each been stuck at 25-35MM in the US for years, while Netflix is approaching 60MM), additional volume and engagement may be necessary. To put it another way, their strategies and incentives differ from Netflix, in part, because their distribution and monetization strategies differ. As the latter converge, so, too, might the former.

To that same end, Netflix’s obsession with engagement may change as OTT video grows from its infancy into a more competitive puberty. As Netflix edges towards domestic saturation, its revenue growth will primarily be driven by price increases – and a reputation for overlong series and B-grade movies may prove problematic regardless of watch time growth (HBO’s price, after all, is 37% higher despite offering a fraction of the library and achieving even less engagement per customer). In addition, the competition in OTT video is only getting stronger. As new entrants attack the space with different priorities, or higher quality thresholds, Netflix will need to respond. Product will not be enough.

Part 1 of this series explained that Netflix spends far more on content than is typically reported. Part 3 will explain why Netflix risks so much.

You can reach Matthew Ball at or @ballmatthew