A few months ago I was talking to some divinity school students, many of whom were on track to being ordained Methodist ministers. I’m always interested in how institutions work, and so I asked them what a successful “career path” looks like in this very un-worldy calling. Amusingly, it turns out the first step after div school looks a lot like applying to any other job. You indicate your interest to a board, and they look at your div school grades, your letters of recommendation, and you go in for an interview. You’re placed at an initial posting, as the years go by you’re evaluated by how well your congregation’s doing, both by size and finances. Inevitably, political matters like how well you play with the church leadership and perhaps any interesting theological publications you’ve put out can also influence your standing. If you do well, you get progressively assigned to larger churches closer to the centers of influence.
And that’s when I realized that nobody goes to Methodist heaven.
Fear not, we intend to be equal-opportunity here! In fact, let’s go nondemoninational and talk about megachurches, which have increasingly become a fixture of the church ecosystem as smaller churches hollow out . These churches can hail from any denomination, or often none at all. There are many reasons for their growth. They’re large enough that you can connect with lots of people in your demographic, which can be a big draw. They’re also often well-run in general. Both the business functions and the production aspect of worship services scale really well, and so the few megachurches I’ve been to have all had excellent sound systems, audiovisual aids, and graphic design. This model has been successful enough that decent-sized cities often have multiple competing megachurches, each with its own spread of services, demographics…and message.
And therein lies the problem. Different amenities are just noise – interchangeable nice-to-haves that don’t much relate to the content of the religion. I’d expect mega-mosques and mega-pagodas in the same city to have similar offerings. Market forces will lead similar groups to trend towards similar practices – towards, say, deciding that having a daycare during services is worthwhile but doing Bingo night doesn’t bring in enough new parishioners to make it worth the cost.
But if market forces can mold churches towards particular amenities, they can mold teaching as well. If people are deciding between different churches, all things being equal they’ll look for a church whose preaching resonates with them, one that they find congenial. In other words, pap. 
It’s important to realize that none of this necessarily involves any ill will on anyone’s part. Normal devout people can disagree in all good faith. The problem lies in having a selection mechanism over this normal range of variation in beliefs. If the market rewards populist preaching with bigger congregations, greater revenues, and heightened influence, then it will be the preachers who preach the popular fare – again, in all good faith – who get outsized influence over the content of the religion. From the next generation, raised with the popular interpretation of the religion, the most palatable preachers among them gets additional weight. Round and round we go, and at the other end of this process: market-optimized pap. And because this process doesn’t require ill intent, “examine in your conscience whether it’s true” and other ways to internally guard against mendacity won’t work. As long as people do vary, there’s enough grist for the mill.
The same considerations apply to the Methodist selection process as well. The grades and politicking aren’t the problem – they may not be ideal criteria for advancement but at least they’re reasonably random. The problem is measuring success in part by how popular and well-received your sermons are and how your congregation grows. Once the market-optimization process starts, your principled beliefs will be worn down to pablum in a few centuries.
One defense against this line of reasoning is to claim that people’s preferences are, if examined, deeply aligned with theological truth. If we can just encourage people to introspect a little, then “what sells” will inevitably correlate with what’s true. Yet this explanation doesn’t hold water on several levels. First, introspection is hard and most decisions about what churches to attend will not be made by outstandingly introspective individuals. So while you can encourage individuals to introspect, it’s impossible to do this for enough people to change the selective landscape. Second, the long record of history shows a broad flowering of many mutually contradictory religions. Popular religions do have some common moral themes and mythic tropes, but they contradict each other in nearly every other theological point. The truth does not contradict itself. So at least most people who have pursued intuitively pleasing theology have been badly mistaken. Finally, the entire claim that people’s intuitions are good theology-detectors is flimsy. If you believe in a God who created the universe, you believe in a God who created quantum mechanics. And you expect His theology to be intuitive and appealing?
Imagine for a moment that we conducted science the way we conduct theology. Every week, we gather at the lecture-hall of one prominent scientist or other who gives a talk on his theories and is supported by donations from attendees. Anything that’s too counterintuitive or has too much math would meet with empty pews. Forget quantum mechanics, I don’t think we’d even have moved past geocentrism! 
What’s needed is a counterbalancing force, not accountable to the people, and charged with maintaining the integrity of the faith. While the elders of a church can theoretically serve this role, it makes it more difficult when they themselves are drawn from the pastors who won the populist competition. To put it puckishly, if you don’t have an Inquisition, get ready to lose your church.
In fact, many people, some of which were even not named Torquemada, recognized that mere intuition would lead people astray, and have attempted to grapple with this problem. In typically sensible fashion, CS Lewis advocated sticking with a parish church (one whose congregants were drawn from a single district) rather than indulging in church-shopping that can fuel competition between churches. If everyone simply went to the local church and never changed their allegiance, then pastors could feel free to preach their interpretation of scripture without worrying about how popular their message would be. He has his demonic Screwtape inquire of an underling:
You mentioned casually in your last letter that the patient [that is, a human that the underling is supposed to be tempting] has continued to attend one church, and one only, since he was converted, and that he is not wholly pleased with it. May I ask what you are about? Why have I no report on the causes of his fidelity to the parish church? Do you realise that unless it is due to indifference it is a very bad thing? Surely you know that if a man can’t be cured of churchgoing, the next best thing is to send him all over the neighbourhood looking for the church that “suits” him until he becomes a taster or connoisseur of churches…
The search for a “suitable” church makes the man a critic where the Enemy [that is, God] wants him to be a pupil. What He wants of the layman in church is an attitude which may, indeed, be critical in the sense of rejecting what is false or unhelpful, but which is wholly uncritical in the sense that it does not appraise—does not waste time in thinking about what it rejects, but lays itself open in uncommenting, humble receptivity to any nourishment that is going.
Another fascinating essay addresses the problem a different way. The author dispenses with primitivism (“The current institution-of-the-Church has been corrupted, oh hey I rediscovered the tenets of the original pure Church”) and sola scriptura (“Don’t rely on old teachings, interpret the scriptures yourself and go with that interpretation”)  as both being absurdly vulnerable to motivated cognition. He then argues – with some impressive Biblical citations – that since the Christian Church was founded by Jesus, who considered the institution a pretty big deal, it’s reasonable to expect that it has a sort of holy oversight that protects it from falling into major errors. It’s not that individual humans in the church won’t err – humans are human – but that given the importance of its mission, the institution of the Church itself is protected from falling into heresy. This is a fascinating argument in favor of If Christian Then Catholic (or Orthodox, I suppose), but it does have one interesting conclusion, which he didn’t draw. If this were the case, you could actually experimentally derive additional bits of theological information! All you need to do is stir up political fights within the Church and see which faction’s interpretation of scripture wins. It would be pretty awesome if the pope started appointing a Chief Experimental Theologian whose job was exactly to set up these sorts of experiments and publish the results. “Blessed are the troublemakers!”
But let’s not rag on Catholics too much. They have, I think, made one of the best attempts at taking a stab at this question. Namely, they’ve arrived at a doctrine called the Communion of Saints. In its more pure-theology aspect this doctrine states that the Church exists outside of time – it’s the community of all Christians including those who died before and the spirits of Christians in purgatory or Heaven. This is why it makes sense to pray for the intercessions of saints in heaven; because they have the backing of an omnipotent deity, they can act throughout time, including right here, right now. But this doctrine has another aspect as well, which is to bind all Christians together by the thread of a common Christianity. A 3rd century monk is as Christian as you, as is your descendant praying in a cathedral on Mars in the 25th century. If your beliefs start to differ from either of them, Bayesian reasoning suggests that you are probably the odd one out, the one who’s drifting into heresy – you’re probably not the enlightened one who finally, at last, found something that everyone else has overlooked. Now, it’s not obvious that this doctrine actually constrains belief that much in practice, and certainly even Catholics have had a lot of doctrinal drift through the centuries. But it’s a very cool idea – the late Romans developed something that looks like Aumann agreement millennia ahead of schedule – and it’s a way to enforce some restriction on doctrinal drift by pointing out that changing beliefs over generations actually is a potential problem and something to be reasonably skeptical about.
Overall, though, this is a really tough problem. It’s really on the order of “What do you do if the Church has been hijacked by demons?” Only in this case, the demons aren’t the dudes with the horns but rather impersonal, hostile optimization processes. And I haven’t come across any really solid exorcism strategies out there. There are a few stabs at ideas, like Lewis’s approach to tamping down selection pressure, and doctrines that enforce Aumann agreement across space and time. But it’s not enough, and worse yet it doesn’t seem like a problem that most religious folk even recognize.
And although this can be turned into a religion-bashing argument, it has disturbing implications for atheists as well. Values matter to everyone. And the fundamental question that these thoughts raise is: how do you ensure that future generations will continue to hold values you consider worthwhile? Besides the winds of popularity, what’s perpetuating your values, keeping them from drifting? If you don’t have a good answer, get ready to lose them.
 In fact, the survival of churches in recent decades have been directly proportional to their size; even among megachurches the biggest have gotten bigger, and even among small churches the smallest have withered away fastest. “For he that has, to him shall be given: and he that has not, from him shall be taken even that which he has.”
 Interestingly when I discussed these concerns with several smart Christian friends, they all instantly seized on the same example of erroneous popular preaching – the Prosperity Gospel. Briefly, this doctrine says that financial well-being is one of the blessings bestowed on devout Christians – if you’re faithful, pray really hard, and donate generously, wealth will be your reward. That this heresy, among all the many popular heresies out there, was universally called out for ridicule, was remarkable. It reminded me of how some atheists use religion as the canonical example of irrationality, even when there are much greater irrationalities in the world.
This whole business smelt of politics to me, and I suspect that the reason the Prosperity Gospel is a popular hobbyhorse among my demographic is that it’s so crass, so very very lower-middle-class – the same differentiating impulse that makes just-barely-middle-class people assiduously avoid Dunkin Donuts in favor of Starbucks, or a young teenager vehemently disparage kids’ stuff. It would be much more awkward for them to call out heresies that are popular and high status among educated folks, such as liberation theology or moralistic therapeutic deism.
 TED talks are pretty much set up along these lines, with exactly the results that we predict.
 I have been informed that this is interpretation, while popular, is not really the true meaning of sola scriptura. The real version makes a much weaker claim – namely that scripture alone is theoretically sufficient to bring a reasonable man unto salvation, and that Church Fathers, tradition, and so on are not strictly necessary. In practice, the content of these authoritative traditions is pretty good and following them is usually the smart thing to do – real sola scriptura looks a lot more like pre-Vatican II Catholicism than modern Protestantism.
By analogy, you can believe that physics can theoretically be rederived from experimental evidence alone but should probably still be mostly reading textbooks and journal articles rather than setting up a lab in your basement to rederive c. In the once-in-a-generation case where an experiment disproves the textbooks, you should go with the evidence – but keep in mind that experimental error and crackpots are much more common than revolutionary discoveries.
This version is significantly less silly and I would only note that it still relies on individual rationality to determine how much to rely on teaching vs. independent thought. In lacking any guidelines except community oversight, it’s still quite vulnerable to decay through pernicious signaling games or just sheer human pride.