Facebook is Building a Subscription Product. Should You Do The Same?

Last week, Facebook/Instagram announced that they are releasing a subscription tier in the EU.

This product will remove ads on the platform and cost €9.99/month on the web or €12.99/month on iOS and Android (the difference likely compensating for the app store fees).

Whether this is a “smart” move is kinda moot point, as the EU is forcing them to do this.

A more interesting question is whether Facebook should expand this program into markets that they aren’t compelled to and what other companies can learn from this.

Why Companies Build Hybrid Subscription Products

There are a lot of companies that have combined a subscription product with another revenue model in a bid to grow their business.

• Spotify, Youtube, DuoLingo - Have an ads business + a subscription product. Buying the subscription product removes the ads.
• NY Times, WSJ, Financial Times - Have an ads business + a subscription product. Buying the subscription unlocks the full product but does not remove the ads
• Uber, Lyft, Doordash - Have core services (delivery and/or rides) + subscription product + ads. Subscription gets you a small discount on the core service but does not remove the ads.

Exactly how these different products combine to raise revenue depends on their business/product.

For Spotify & YouTube, their free product is expensive to run, driven by the cost of music royalties and video hosting respectively. They need the ads to offset the cost and (hopefully) make these products profitable.

Having ads allows them to make the free product sustainable and these ads are annoying enough (because they take 100% of your attention) that people will pay to remove them.

For the NY Times, WSJ, and Financial Times, they have figured out that paying subscribers are still going to use the product even with ads, so they don’t have to remove them. Also, their ads are shown next to the content, so can be more easily ignored.

For Uber/Lyft/Doordash and other similar companies, the membership products are defensive moves, with the theory that subscribing to UberOne will make you more committed to Uber and less likely to switch to a competitor.

Additionally, these are low-margin businesses, so the high-margin nature of ads helps their bottom line.

For Facebook, because of the regulatory requirements, the subscription tier allows them to keep operating in these markets and not lose a big chunk of their revenue base.

How Might this be a Good Thing for Facebook?

Besides being compelled to do this to keep operating in Europe, let's look at how this might go well.

1. Recurring revenue brings predictably

Recurring revenue products are significantly more predictable than ads which are more directly influenced by external factors (interest rates, political elections, fundraising cycles, browser/device tracking policies, etc).

Recurring revenue businesses, though not without challenges, align internal incentives more directly with your users. Keep making a great product, the users pay you directly.

Despite that each individual payment is much smaller, e.g. €9.99 monthly from a consumer vs €600k annually from a small CPG brand, the consumer payment amounts don’t change whereas ad spending fluctuates.

Entering what could be uncertain economic times in the next few years, having a secondary and more stable revenue stream will help add some predictability to the business.

2. Recurring revenue brings durability

Ad-based businesses are much more prone to the “influence” of those advertisers and the need to keep the content on the platform compliant with the differing standards of those advertisers.

All content-based platforms have to content with larger advertisers trying to influence what content they appear next to and they always make the threat of dropping spending or pulling out entirely. This problem can never fully be solved.

The durability of recurring revenue can’t be understated. In addition to diversification of the revenue base, meaning you have a lot of little payments that make up your gross bookings vs a few large clients, people can stay subscribed for a long, long time.

In 2021, there are still 1.5 million people paying for AOL mail, which was released in 1993 and became free in free in 2006.

If you’re tackling a never-ending use case (which Facebook is) and very good at payment processing, then these revenue streams keep going for a long, long time.

3. It (slightly) helps counter the narrative of Facebook “selling your data”

While this is by far the least immediately impactful benefit, this does fly in the opposite direction of Facebook’s biggest criticism, that they "collect all your data and sell it".

The public has this perception of Facebook despite that:

1. This isn’t really how advertising works
2. Consumers have clearly signaled that they are fine with it
3. EU regulators forced them to make this change

However, offering a subscription tier is still helpful to the company's PR efforts, relationships with local governments, and overall narrative going forward.

Additionally, each country in the world is building its version of GDPR, and privacy laws as a whole are getting a lot stricter, so diversifying away from a pure ads business is a good idea.

How Might This Be a Bad Thing For Facebook?

1. It removes more agility from the business, especially around pricing

In addition to adding more complexity to the business overall, adding a subscription tier with public-facing prices has significantly less agility than ads.

The price of an ad is set by the Facebook ad auction, which is influenced by a number of factors some of which are in Facebook's direct control and some aren't.

Facebook is incentivized to keep prices reasonable so that they keep the ROI for advertisers, which keeps them coming back.

That said, if Facebook wants to make their auction more expensive, they can do so behind the scenes, without a press release, and without most people knowing about it.

Once you have public-facing pricing for a subscription product, changing that price is a lot more complex and risky.

Facebook will have done the math on what they need to price their subscription at in order to compensate for the loss in ad revenue per user. However, if they were wrong, this is much harder and slower to fix.

2. Making tradeoff decisions becomes much harder, which slows product development

In a giant tech company, there is a heavy reliance on A/B testing to make decisions. This has a lot of advantages:

• You have tons of traffic, so you can measure changes down to very small increments and still get statistical significance
• Excellent A/B testing and measurement systems, so you can effectively track everything
• Really, really strong data scientists who can find you an answer to basically any question

However, A/B testing gets very complicated when you have to compare tradeoffs between different products that might not be comparable.

When I was at Uber Eats, we shipped everything (including infrastructure upgrades) as an A/B test. This worked really well when we were just trying to grow the food delivery business.

However, as the business evolved, we had to consider our impact across the range of products. Each of which we had different levels of understanding & historical data.

We had to think about:

• Restaurant Delivery - The core businesses. We had a ton of data on usage, average order size, expected retention, etc
• Grocery Delivery - Growing business, totally different usage pattern, lower margin, much less historical data
• Restaurant Ads - Very high margin, tough to understand the negative impact on the platform. Big company priority.
• Uber One Memberships - Early in its development, LTV is not really yet understood.

Because we can measure changes down to 0.001% level, each time we run an experiment you can see multiple metrics move in different directions across the different products.

This means that you will A/B test a feature that:

• Helps restaurant delivery
• Hurts grocery delivery
• Helps ads
• Hurts memberships

Now you have to decide if this is something you should release to 100% of the users.

Luckily Uber had excellent data scientists who could develop tradeoff criteria for these metrics, but the point is that this really slows down product development.

Every A/B test needs to run longer and be debated longer in order to come to a decision.

So What Should You Do With This Information?

While I (clearly) love subscription products, I don't they should be combined with other business models unless a company absolutely has to.

The additional complexity and difficulty you'll have compared them with shorter revenue lifecycle products (like ads) will dramatically slow things down.

That said, once you have to build one. Follow the best practices and be patient.