This post was originally written for Recode.

Facebook has a problem. What has driven its growth for the last five years won’t drive its growth for the next five. However, the options in front of the company involve the kind of user experience compromises that have maimed platforms that preceded it.

Facebook makes its money from the West. Some 30 percent of its users and 73 percent of its revenue is from North America and Europe. The monthly average revenue per user for Western users is $3.33 versus 53 cents for the rest of the world. Facebook is a global company, but a Western business.

Facebook’s user growth in the West is a little over 1 percent a quarter. In North America, Facebook’s monthly active users represent 80 percent of the population above the age of 14. If Facebook wishes to grow its Western revenue at the rate its shareholders demand, a 1 percent user growth rate will not do it.

Absent rapid user growth, the other lever for increasing advertising revenue is increasing the number or value of ads that are shown to existing users. However, the News Feed is close to saturation. Facebook believes that it cannot stick any more ads in the News Feed without adversely affecting user retention.

This combination of slowing user growth and News Feed saturation has led Facebook to warn of a rapid deceleration in revenue growth over the next six months. For the first time in years, Facebook needs a new lever to pull. However, its options largely involve moving from the kind of advertising that most people are fine with, to the kind of advertising that most people are not.

Effective advertising is about picking your moment. Facebook’s News Feed ad units work because they interact with consumers at the moment of choice. Users scroll through a sequence of content and commercial placements as they decide what to consume next. Ads at the moment of choice, such as full-page ads in Vogue, paid links in Google or T-brand studio-style native are effective and generally acceptable to consumers. The trouble begins when you have to move past the moment of choice and advertise against the content itself.

Once past the moment of choice, one can either advertise adjacent to the focus of consumption (no longer an option on mobile) or create advertising that interrupts the consumer from doing what they wanted to do. This is the world of pre-rolls, interstitials, auto-expanding video units and takeovers that have made the mobile web such an unusable eyesore.

Facebook has traditionally avoided trying to interrupt consumption after the moment of choice. There are no interstitial ad units when browsing a family photo gallery, no calls to action in the comments on your friend’s post. However, a saturated News Feed means it doesn’t have that luxury. Facebook faces a Sophie’s choice: Insert inventory within the user-generated content it controls and tempt the fate of Myspace, or make already ad-supported publisher inventory its own and deal with a host of fractious partners scrabbling for revenue.

Facebook has chosen the latter path with its two main attempts to expand beyond News Feed ad units: Video and Instant Articles. Facebook’s hope was that it could find a better compromise between revenue and UX than publishers have before it. Instant Articles could be faster with a more acceptable ad touch than the mobile web; video ads could be mid-roll rather than pre-roll. However, for the first time Facebook is dealing with a shared revenue model that must be attractive enough to woo publishers. As a result, its desire to avoid hurting long-term user engagement has sprinted headfirst into the brick wall of publisher revenue-share expectations.

Instant Articles began with an ad every 350 words before compromising to 250 words, recently adding even more ads in the recommended links at the bottom. The ads themselves went from click-to-play video to silent autoplay, and from silent autoplay there is commercial pressure to move Instant Article ads to autoplay with sound on.

Even with that, premium publishers can’t make the revenue math work, and premium uptake has been insipid. Facebook’s best option is to stop focusing on top-tier publishers whose revenues it cannot match, and instead focus its efforts on the long tail of publishers and independents whose revenues are easier to beat. It takes away the premium brand story Facebook would love to sell to advertisers but at least drives inventory expansion.

Video is at an earlier stage, but is following a similar path. Facebook is experimenting with mid-roll ads. Mid-rolls enable more video engagement but get seen less than pre-rolls. One publisher generated $11,000 from 24 million video views, a $0.46 eCPM that is an order of magnitude less revenue than the same views would generate on pre-roll-friendly YouTube. Compromising the user experience to pre-roll seems the obvious move, but the News Feed’s dynamics make it far harder than one might think.

Unlike intent-driven interfaces like YouTube, the News Feed is a passive experience, where it is easier to scroll than to engage. A News Feed consisting of an endless stream of autoplaying pre-roll ads would be an engagement nightmare that would compromise everything the News Feed team holds dear. Lacking the flexibility to compromise, they must give elsewhere. In February, Facebook embraced sound-on autoplay videos, just months before Google and Apple said they were banning these formats on their platforms as unacceptable UX intrusions.

Facebook isn’t ignoring the problem. It has built out a separate video tab to try and drive the kind of intent-based consumption that might be more open to pre-rolls than News Feed. However, to get consumers to engage requires a strong bank of high-quality video content that it does not have. Moreover, with monetization weak, the incentives for publishers to create that content of their own volition are not there.

Thus, Facebook is priming the pump by commissioning video or offering guarantees so that the opportunity-cost calculation is removed. However, in doing so, Facebook is embracing paid content creation, the avoidance of which has in part been key to its success. This is a risky endeavor that depresses margins, and where the hits can be mercurial. Just as with Live, there are no signs that Facebook wishes to be in this game long-term.

Facebook is stuck between a rock and a hard place. As much as Facebook wants to create a better experience than the mobile web it disdains, the pressure to make these new inventory opportunities big enough and fast enough forces it down a road of increasing UX compromises by publishers who long since fought the war between short-term revenue and long-term user engagement, and too often chose the quarter over the future.

The very real fear for Facebook is that the publisher UX compromises it has watched from afar are not a path it has managed to avoid, but a vision of its future.