The desert moved to the growth edge

For twenty years the picture of a gospel desert in this country has been an urban one. A block of vacant storefronts, a church building sold and turned into apartments, a neighborhood the denomination left in 1974 and never came back to. That picture is not wrong. It is just no longer where the numbers point hardest.

Three separate questions were asked of data that already existed. The first asks where the ground is measurably getting better while gospel access stays worst-tier. The second asks what any of this looks like when you stop describing places by eleven-digit census numbers and start using the names people actually say. The third asks which kinds of households stay unreached, and whether they are unreached in the same way.

The three were run independently, by different joins, over different columns. All three land on the same ground: subdivisions, exurbs and university towns that were fields fifteen years ago and now hold ten thousand households and one church.

The Gospel Desert Index is a standing level. It scores structural gospel access from nought to a hundred, where higher means less access, and it counts congregations, their distance from the tract line, and the conditions of the households around them. It says how things are. It says nothing about direction.

The Gospel Impact Metric is the direction. Its newer long-baseline form fits a trend across twelve overlapping five-year windows rather than differencing two eras, which cuts the noise by roughly two thirds. Its most important feature is the one that makes it less exciting: it will decline to answer. When the fit cannot separate rising from falling, it reports holding, and holding is not a finding that the tract has held steady. It is a statement that the direction is not established. Twenty-six thousand tracts sit there, and they are excluded from everything below.

The Missional Viability Index asks a question neither of the others can: whether there is anywhere to meet. Third places, street intersections, public rooms. A tract can be opening and unreached and still be a place where nobody can gather anybody, because there is no room and no corner and no reason to stop walking.

Cross the three and you get a claim none of them makes alone. The ground is improving. Access is still among the worst in the country. And the built form could hold a congregation if one existed.

Across the analysis universe, 38,849 tracts are confirmed opening on the long baseline. That is a large number and on its own it is not very useful; most of those tracts already have churches, or are comfortable, or both. Hold them to Arid or Desert access and an MVI above the midpoint, and 2,168 remain.

Then read the list.

Jacksonville is on the list, but not the Jacksonville anyone pictures: this is a tract on its southern build-out, eleven congregations, access still Arid, and the fastest confirmed improvement in the country at that access level. Flagstaff and Auburn and Oxford are university towns whose edges filled in. La Vergne is a Nashville commuter suburb whose tract holds no congregation at all and reads 74 on access, the worst figure on the page. Mascotte and Davenport sit on the Orlando build-out. Greer is on the I-85 corridor between Greenville and Spartanburg.

Now read the last column, which is the one that matters. Working New-Subdivision Families. Affluent New-Build Families. Exurban Working Homeowners. Cost-Pressed Family Suburbs. These are not the household types anyone means by the phrase unreached. They are the household types that show up in a church growth seminar as the target audience.

That is the problem. A subdivision built in 2014 has no church in it because nothing was in it. There was no congregation to leave and no denomination to blame. Ten thousand households arrived over eight years into a place with a road, a school and a grocery store, and the ordinary mechanism by which a neighborhood acquires a church, which is somebody already being there, never ran.

What it asks of whoever goes

The household guide for the type that tops this list opens with a sentence worth reading twice: four in five of these houses went up since 2000, so there is no history here, no institutions, and nobody older to ask. That is the pastoral situation in one line, and it is not the situation most training prepares anyone for.

Consider what is absent. There is no matriarch who knows everyone, no funeral home that has buried three generations, no high school with alumni still nearby, no lapsed member who could be visited, and no building with a name on it. A planter arriving in an old neighborhood inherits a set of relationships, most of them complicated; a planter arriving here inherits nothing at all, including the complications.

That cuts both ways and the good half is real. Nothing has to be repaired. Nobody remembers the last church that failed here, because there was not one. Everybody on the street is also new, which is the single condition under which a stranger is not a stranger. It does not last. The window in which a whole neighborhood is simultaneously looking for people closes within a few years of build-out, and it closes quietly.

The bad half is that the ordinary tools do not work. There is no crowd to join, because there is no place where this neighborhood is ever gathered. The guide names the one exception, and it is worth taking literally: the pickup line at three o'clock. Fifteen minutes standing there weekly beats any invitation anyone could send, because it is the only time these households are out of their cars and in the same place.

And the economics are misread almost every time. A median household income near a hundred thousand dollars, in a place like this, is servicing a new mortgage, two cars, childcare and a fifty-minute commute. The guide's warning is blunt: upper-middle income here does not mean slack. Two commutes and small children squeeze the usable evening down to about ninety minutes. A church that schedules a Wednesday programme at seven is not competing with apathy. It is competing with the drive home.

The practical consequences are unglamorous and specific. Move everything to half past eight in the evening or to Saturday morning. Pick one gathering on one date and hold it, unchanged, for three years, because a place with no institutions cannot tell the difference between a church and a pop-up until something has survived long enough to be predictable. Find the households who arrived in phase one and give them real information to pass on, because in a neighborhood with no memory the earliest residents are the closest thing to elders it has.

And do not admire the commute. The guide is precise about why: calling a man's fifty-minute drive heroic makes it harder for him to say that it is grinding him down.

There are two ways to ask whether a tract is improving. Difference the last era against the one before it, or fit a line across twelve overlapping windows. They are different questions, and on 846 of the 2,168 opening deserts they give opposite answers.

Flagstaff reads −2.9 on the two-era delta and +14.8 on the long baseline. Both are correct. A tract that fell for six years and then recovered for three produces that pattern: the recent pair looks bad, the long line looks good.

The temptation is to pick one and call it the real number. The rule this project works under says otherwise, and it is the right rule: where the two disagree, the disagreement is the finding. A place that fell hard and is now climbing is a different pastoral situation from one that has been climbing steadily, and flattening them into a single figure would delete the difference that matters most to whoever is sent there.

Every figure above is attached to an eleven-digit census tract number, which is a fine thing for a database and a useless thing for a person. A pastor does not carry a burden for 24003706302. He carries one for Eastport, which is the part of Annapolis where that tract is.

Naming the ground turned out to be hard, and the reason is worth a paragraph. The federal gazetteer drops about eight named places into an average tract, and most of them are subdivision plat filings: the name on the sign at the entrance to a development, filed once in 1987 and known to nobody. Four separate ways of ranking those names were tried against Annapolis to find the real neighborhood among them: nearest to the tract centre, encyclopedia notability, whether a Wikipedia article exists, and the population living around each point. All four preferred Watergate Village and Colonial Manor Estates to Eastport.

The answer was to stop choosing. Where a city publishes its own neighborhood boundaries, use those. Where it does not, use what contributors have mapped, because people map the names they say out loud. Only then does a ranking of neighborhoods mean anything.

Rolled up that way and weighted by households: 6,765 named areas carrying three tracts or more, of which 1,425 are neighborhoods rather than whole towns.

The neighborhood list still reads like the old picture. Warrendale in Detroit, Carver Shores in Orlando and Central Southwest in Houston are the places a denomination would name if asked to guess. But widen the same query from neighborhoods to every named area and the top of the list changes character. Poinciana, Florida is the largest of the driest: eleven tracts, 21,527 households, forty-one churches, and a household-weighted access score of 74.9. Beside it sit Interlachen and Davenport in Florida, Knik-Fairview in Alaska, Lake Los Angeles in California, Chaparral in New Mexico, Anderson Creek in North Carolina.

Every one of them is a place that got built quickly. Several did not exist as named places within living memory.

Why the name is a pastoral question and not a formatting one

It would be easy to file the naming work under presentation. It is not. A tract number cannot be prayed for, cannot be preached about, cannot be put in a bulletin, and cannot be the thing a person says when asked where they are from. Every act of ministry that this data is meant to serve happens under a name.

There is a second-order problem, too, which the growth edge makes sharp. A neighborhood that has been named for a century has a name that means something: Eastport, Southmost, Warrendale. A subdivision has a name a developer chose to sell houses. Dunhill Meadows, Ironwood Terrace, Anthem Highlands: those names arrive before any of the shared life that usually produces one. Both kinds appear in the table above.

That is not an argument for dismissing the developer's name. People do live in Anthem Highlands and they do say so. It is an argument for noticing what a pastor is walking into: in one case a name carries a hundred years of meaning and every possible offence, and in the other it carries a brochure. The work of turning the second into the first is most of what planting on the growth edge is.

The third crossing asks which of the thirty household guides stay unreached. For each guide: its share of households living in unreached tracts, divided by its share nationally. A figure above one means the type is over-represented among the unreached; below one, under-represented. Weighted by the full composition of every tract rather than its dominant type, so a tract that is sixty-forty counts as both.

The important move is splitting unreached into two situations that call for opposite responses. Beside a church means a congregation sits inside the tract and access is still Dry or worse. None inside means there is no congregation at all. Twenty-four thousand tracts are the first. Forty-seven hundred are the second.

Look at High-Achievement Homeowners. Among tracts unreached beside a church they run at 0.54× their national share, under-represented. On that column alone a denomination would conclude that its affluent professional households are covered, and move the money elsewhere.

The second column reads 1.62×.

They are not reached. They are over-represented in tracts where no congregation exists at all, so they barely enter the first bucket in the first place. Their scarcity among the misaligned is not evidence of success; it is a consequence of there being nothing there to be misaligned with. Transit-Dependent Gateway Households show the same inversion more sharply still, at 0.59 beside a church and 1.78 with none.

Affluent Transit-City Creatives make the point from the other end. They read 0.10× beside a church, the lowest figure in the set by a wide margin, and 0.88× with none. That second figure is close to par. They are not nine times better reached than the average American household; they are roughly average, and the first column overstates it by about that factor.

That is a methodological point rather than a finding, and it is the one worth carrying out of this article. Any one of these tables, read alone, will tell you something confidently untrue.

Read both columns together and the shape resolves. The types over-represented among both kinds of unreached are the growth-edge households: new-subdivision families at 1.80 and 1.72, affluent new-build at 1.45 and 1.98, car-oriented gateway families at 1.40 and 1.89. Not unreached because a church left. Unreached because a church never arrived.

And only three guides out of thirty are scarce in both columns: Highway Corridor Commuters, Middle-Income Established Homeowners and Young Multifamily Workers. Those three are the only household types this data is prepared to call reached.

None of this measures faith, and it should not be read as though it does. It measures access, meaning whether a congregation is near enough and stable enough to be reachable by the people living there, and it measures the movement of the conditions those people live under. A tract can be full of believers and score badly. A tract can score well and be cold.

What the numbers can do is tell you where to look, and stop you looking only where you already were.