Gallup's March 2026 measurement puts weekly or near-weekly attendance at 31%, with 10% monthly and 57% seldom or never. Religion rated "very important" sits at 47%, down from 58% in 2012. Americans reporting no religious affiliation reached 24%, the highest on record, against 2% in 1948.
Pew's 2023-24 Religious Landscape Study, published February 2025, found Christians at 62% of U.S. adults and, in the underreported finding, the decline has stalled. The Christian share has been "relatively stable since 2019." Monthly-plus attendance has hovered in the low 30s since 2020. Pew also found that 35% of U.S. adults have switched religion since childhood. More than a third of the market is in motion.
Then the finding that should redirect a lot of budgets. Gallup, April 2026: young men calling religion "very important" jumped to 42%, up from 28%, back to their turn-of-the-century peak. Young women held flat at 29%. On monthly-plus attendance, young men reached 40% and young women 39%, statistically tied, closing a gap that once ran sixteen points in women's favor.
A caution worth printing. Barna and Gloo report considerably larger and faster shifts than Gallup or Pew: attendance rising to 32% in early 2025 from 28% in 2024, and a 43%-male-versus-36%-female weekly gap they call the largest in twenty-five years of their tracking. Political scientist Ryan Burge has publicly disputed the revival narrative, and the American Institute for Boys and Men has pushed back on the young-men figures. Barna's methodology uses quota sampling with minimal weighting, while Gallup and Pew use probability samples. Attribute Barna numbers to Barna and do not blend them with Gallup or Pew. The direction of the gender finding is independently corroborated. The magnitude is not.
Strategy one: how should a church website be built?
Build it for the visitor, not the member. The most consequential church marketing failure is measurable and almost universal.
A 2023 UX study analyzing 50-plus church websites across congregations from 600 to 45,000 weekly attendance found the top navigation items in this order: Give appears on 86% of sites. Watch Sermons 80%. About Us 76%. Locations 40%. Connect 38%. Events 34%.
Read that as a priority statement, because that is what it is. Giving is on more than twice as many navigation bars as location. The site is built for the person already in the building.
What first-time visitors look for has been stable for over a decade: service times, address, which campus, what to expect, children's ministry, who the pastor is, what the church believes, and accessibility. A 2012 study of 1,011 adults found service times the single most common reason for visiting a church website at 43%, ahead of activities, directions and streaming. The 2023 UX research corroborates the ranking.
The fix is not a redesign. It is a re-ordering.
- Put service times and address above the fold on the homepage. Not on a Visit page. On the homepage, in text, not inside an image.
- Cut main navigation to three to five items. Six maximum. The 2023 research is explicit on this, and church sites routinely run nine or more.
- Build one honest "what to expect" page. How long is the service? What do people wear? Where do I park? What happens to my kid? Is there coffee? Will anyone make me stand up? These are the actual questions, and most church sites answer none of them.
- Fix your denominational listing. Grey Matter Research audited roughly 8,000 listings across 123 Protestant denominations and found a 32% chance of successfully finding a local church through a denomination's own website. Half the listings lacked a web link even when the church had a site, and one in five links was broken. Check yours this week.
Strategy two: should churches livestream?
Yes. The cannibalization fear is wrong, and this is the strongest single finding in church research right now.
The EPIC 2024 Attender Survey, with over 24,000 churchgoers across 80-plus denominations and published in June 2025, states it directly: "Offering a live stream does not reduce in-person attendance, nor does it erode the importance of belonging to a faith community."
The breakdown: 26% regularly engage in online worship. 19% move between online and in-person, the hybrid switchers. Only 4% attend online exclusively, and that 4% is the group showing lower engagement and giving. The hybrid switchers do not. Eighty-five percent of churchgoers belong to a congregation offering some form of online worship, and 86% of churches livestream.
Behavior during online worship is worth knowing before you produce for it: over 60% multitask, 95% pray or meditate along, 71% sing along, and 87% report satisfaction.
Then the finding that should reframe how churches think about their audience. 46% of churchgoers regularly attend or view services at more than one church, physically, virtually or on television. Seven percent claim more than one "home" church.
Church attendance is now a portfolio, not an exclusive relationship. The competitive question is not whether someone leaves. It is what share of their attention you hold, and whether your content is good enough to stay in the rotation. That is a communications problem, and it is measurable.
Strategy three: why turn on recurring giving?
Recurring giving is the clearest ROI line in the entire category.
Digital now accounts for 42% of total church contributions. Recurring gifts make up 42% of digital dollars and 57% of digital transactions. And the operative finding: churches offering recurring giving are twice as likely to see overall giving increase, yet nearly half have not adopted it.
Digital wallet contributions such as Apple Pay and Venmo more than doubled in 2024, while only 20% to 33% of churches accept them. Younger donors are twice as likely to give when the option exists. Cash and check volume held steady, which is the point: digital giving is additive, not cannibalistic.
These figures come from a vendor-published report and the methodology is not disclosed, so treat the magnitudes as directional. The structural claim, that recurring beats one-time and the setup is a one-afternoon task, is not seriously contested.
On the platform side, the Barna and Pushpay State of Church Technology 2026 report, fielded November to December 2025 across 1,306 church leaders, found social media adoption at 97%, livestreaming 86%, QR codes 82% and named mobile church apps the top future investment priority at 59%. Roughly half the sample came from Pushpay's own customer list, so adoption rates likely run high against the national picture.
Strategy four: how does a church win local discovery?
It does so with free tools, and the inventory is nearly free. The median U.S. congregation operates on $205,000 in annual income. Staff compensation typically consumes 45% to 55% of a church budget, facilities another 20% to 30%, with ministries and outreach, where marketing sits when it is tracked at all, at 10% to 15%.
Run that math and the median church has low four figures a year for communications. Any "churches should spend 2 to 5% on marketing" benchmark is a consultant's rule of thumb. No primary research isolates church marketing spend. Not ECFA, not the National Study of Congregations' Economic Practices, not Faith Communities Today, not Lifeway.
Which makes the free inventory the whole game.
Google Ad Grants provides qualifying nonprofits and churches up to $40,000 per month in in-kind Google Ads. That is the single largest real number in church marketing, and it is not cash spend. Most churches never apply.
Google Business Profile is the other one. Accurate service times, correct address, current photos, responded reviews, structured event listings and consistent name and contact data across every directory that lists you are the signals that surface a church in local discovery. They cost nothing but attention.
Sustained beats occasional. A church posting three times a week for a year outperforms one that runs a four-week campaign and goes quiet, because unpredictable cadence trains audiences to disengage.
Strategy five: why publish text?
Machines read text. A structural shift is arriving in local discovery, and church communications is not ready for it.
Generative AI web visits grew 70% year over year to 9.5 billion monthly, with unique visitors up 57% to 655 million. Platform share moved hard in a single year: ChatGPT from roughly 76% to 53%, Google Gemini from under 9% to about 27%, Claude from 2% to 9%. ChatGPT's citation rate, how often answers cite sources, rose from 1.6% in June 2025 to 6.8% in May 2026.
Be precise about what this does not prove. There is no research measuring whether people ask AI engines for church recommendations. Every "AI search for churches" figure in circulation comes from an agency selling the service. Citation rates are strongly category-dependent: travel runs about 23%, automotive about 20%, and professional services under 4%. Local religious organizations most plausibly sit near the bottom of that range.
So the honest position: this is early, and it is cheap to be early. The work is the same work that helps human seekers, and EPR has the mechanics documented in GEO for Churches and Schema Markup for Nonprofits:
- Publish sermon transcripts as structured text, tagged by topic. Text is searchable, retrievable and reusable. Video is none of those things to a machine. A church with ten years of sermons and no transcripts has ten years of invisible content.
- Answer the real questions in dedicated pages. Not doctrinal essays, but the actual questions people bring: anxiety, grief, marriage under strain, parenting, doubt, forgiveness, money. One clear page per question, with a real answer. These are the pages an engine can retrieve and cite, and they are the pages a seeker at 1 a.m. needs.
- Fix your Wikipedia entry if you have one. Most large congregations have outdated entries. Wikipedia sits fourth among all domains cited in AI-generated answers and first for ChatGPT specifically, and the full mechanism is in EPR's Wikipedia and AI citation guide. Accuracy there propagates everywhere.
- Name things specifically. Entity-rich language, with real place names, program names, times and people, is what a retrieval system can anchor to. "Sunday services at 9 and 11 a.m. at 400 Main Street in Birmingham" is retrievable. "Join us for worship" is not.
- Pursue earned coverage in faith-trusted outlets. Christianity Today, Religion News Service and the religion desks at national outlets reach audiences no church website will, and they function as retrieval anchors, the authoritative sources engines weight when generating an answer. EPR's church PR guide covers how a congregation earns local and religious press.
What do the largest church programs actually do?
The Outreach 100 list of America's largest churches is led by Life.Church in Edmond, Oklahoma, followed by Church of the Highlands in Birmingham, Christ's Church of the Valley in Peoria, Lakewood Church in Houston and North Point Ministries in Alpharetta. Two of them are worth studying for reasons that have nothing to do with size.
Life.Church built the infrastructure everyone else uses. Its YouVersion Bible App passed one billion device installs in October 2025, with installs up 12% year over year, daily usage up 18%, one billion app opens every 39 days, 2,300-plus languages, and over 31,000 partners including more than 25,000 churches. Life.Church also gives away its resources through the Open Network and its Church Online Platform. That is the most durable authority position in the category, and it was built by making other churches successful.
Church of the Highlands runs a structured onboarding program, the Growth Track, marketed consistently across web, social and in-service. Converting a visitor into a member is a designed funnel, not a hope.
One cautionary case, stated factually: Hillsong Church has gone from 16 U.S. campuses to 5, with the San Francisco campus separating in February 2025. The remaining U.S. locations are Los Angeles, Orange County, New York City, New Jersey and Boston. Whatever the underlying causes, the communications lesson applies to every organization: reputation infrastructure has to be built before the crisis, not during it. AI engines retrieve the record permanently and summarize it into every relevant answer.
What do weaker church programs get wrong?
Four failure modes, all fixable inside a quarter.
- The site is organized around the org chart. Staff pages, ministry lists and the giving portal, structured the way the church thinks about itself rather than the way a stranger searches. The 86%-Give-versus-40%-Locations gap is this failure quantified.
- Cadence collapses. Two weeks of heavy posting, two months of silence. Audiences read inconsistency as inactivity.
- Content serves the producer. Long theological essays published for an audience asking how to handle a marriage under strain, and video sermons shipped with no transcript, no summary and no accessible alternative.
- First visit is left to chance. The hardest moment in church attendance is the first one, from parking through greeting through service through the exit. Programs that design it deliberately retain visitors that programs relying on friendliness lose.
The bottom line: the difference between the congregations that grew and those that shrank is not the culture. It is whether the website answers a stranger's questions, whether the stream runs, whether recurring giving is switched on, whether the local listings are accurate and whether anyone published text a machine can read. None of that requires a budget. All of it requires deciding that communications is a discipline and not an afterthought.
