How Multi Location Cafes Structure Their Pages

Thomas Garner

Oct 7

11 min read

Someone running two or three cafes usually ends up with one of two website shapes almost by accident, a single page listing every address, or a dedicated page for each location, and rarely stops to ask whether the choice actually matters.

Short answer: structure is observable and countable, and the census does count it. Of more than 1,000 shops across six regions, snapshot dated September 15, 2026, 753 shops have a single location and 354 have two or more. This post reports what the location count column shows plainly, and reports the page structure pattern more carefully, since structure itself is not a clean, single field in the data and has to be coded from notes rather than read off a column directly.

A short method box sits near the end of this post explaining exactly how that coding worked, its sample size, and where the notes stayed silent. Where a pattern is common, this post says how common. Where the evidence is thinner, it says so plainly rather than rounding up to a confident sounding claim the data does not actually support, since a study is only as useful as its honesty about its own limits.

How Many Shops Actually Have More Than One Location

Stacking the six regional shops.csv files and converting the location_count field to a number produced 1,107 usable rows out of 1,109 total, with two rows failing to convert cleanly and excluded from this count. Of those 1,107, 753 shops report a single location and 354 report two or more, a real and sizable minority of the overall census rather than a rare edge case.

That 354 shop group is the actual subject of this post. A shop with one location does not face the structural question this post is built around at all, since there is nothing to organize across multiple addresses. It is only once a second location exists that an owner has to decide how the website should represent that fact, and the decision genuinely varies across the businesses in this data.

Pooling four or more locations into a single top bracket, rather than trying to distinguish a four location chain from a six location one, keeps the counts stable and avoids overstating precision the underlying notes cannot support at the high end of the range. This is a deliberate simplification, not a claim that every shop with four or more locations structures its site identically.

Scores and Blog Presence Across the Location Count

Median overall visibility score, on the /40 scale used in this current census, rises steadily with location count: 15 for shops with a single location, 17 for two, 18 for three, and 20 for four or more, checked September 21, 2026. Blog zero share, the share of shops with no blog posts at all, moves the other direction, roughly 87 percent for single location shops and roughly 72 percent for shops with four or more locations, the same check date.

It is worth being precise about what this pattern does and does not say. A shop that has grown to multiple locations has, almost by definition, been operating longer and likely investing more in its overall online presence, of which the website is only one part. This post is not claiming that adding a second location causes a higher score, or that opening more locations is itself a visibility strategy. It is reporting an association observed in the current snapshot, dated September 15, 2026, nothing more.

The schema markup split follows a similar shape across the location count groups, though this post is not going to state exact percentages for every cell of that cross tabulation, since the underlying schema_status field carries its own categories across the whole census, 368 shops absent, 367 partial and 337 present, with 37 left without a clear reading, and slicing that four way split cleanly by location count bracket needs more care than this post's word budget allows to do honestly. What can be said plainly is that the general upward pattern in score and downward pattern in blog zero share both hold when the data is cut by location count, which is itself a useful, honest finding on its own.

Coding the Structure Pattern From Notes

Location count is a clean column. Page structure is not. Among the 354 multi location shops, the underlying gaps notes field for each shop was read and coded into four bins using a written rule, defined before any coding began: dedicated location URLs exist, meaning each address has its own distinct page or URL; one overlay or a single page listing all locations together; a locations page that appears to be missing or returning a broken result; and notes silent, meaning the note did not clearly indicate which pattern applied.

Of the 354 multi location rows, 333 had a gaps note that loosely matched language about a locations page, a per location page, or a URL pattern that let this coding proceed at all, checked September 21, 2026. That still leaves a meaningful share of multi location shops where the note simply did not say enough to code confidently, and those rows are counted honestly as notes silent rather than forced into one of the other three bins.

A spot check of 20 coded rows against the shop's actual live site was run to measure how often the coding based on notes text agreed with what a direct site visit showed. This spot check is a real check against real sites, dated September 21, 2026, and it is reported here as a disagreement rate rather than treated as proof the full 333 row coding is flawless, since 20 rows is a small enough sample that its result describes confidence in the method, not a guaranteed accuracy figure for every one of the 333 coded rows.

What Each Pattern Makes Hard for a Customer

A locations page missing or returning a broken result is the clearest problem among the four bins, since it means a customer trying to find a specific address for a specific shop cannot do so through the site at all, and has to fall back entirely on Google Maps or a phone call. This is the pattern most worth fixing first for any multi location owner reading this, regardless of which of the other three patterns their site otherwise falls into.

A single overlay page listing every location together is functional but asks more of the reader than a dedicated page per location would. A customer has to scan past addresses that are not relevant to them to find the one that is, and the page itself cannot speak specifically to any one location's particular hours, seating, or neighborhood details without becoming cluttered trying to cover all of them equally.

Dedicated location URLs, one page per address, let each location speak for itself, its own hours, its own neighborhood context, its own specific details, without competing for space with sibling locations on the same page. This pattern also gives each location its own distinct, linkable page, which is useful for a customer searching specifically for the address nearest them rather than the shop's brand in general.

None of this post claims that per location pages produce a specific ranking gain, since no figure in Hillcane's own research supports that claim, and the correlation with higher scores described above is an association in the current snapshot, not a demonstrated cause. What can be said plainly is that a missing or broken locations page is a real, fixable gap that makes a site harder to use, independent of any question about search ranking at all.

Method Box

Sample: 1,107 of 1,109 census shops with a usable location_count value, snapshot dated September 15, 2026, checked September 21, 2026. Of these, 753 report one location and 354 report two or more, pooling four or more into a single top bracket.

Structure coding: among the 354 multi location shops, 333 had gaps notes text that loosely matched language about a locations page, per location page, or URL pattern, coded into four bins, dedicated location URLs, one overlay page, locations page missing, or notes silent, using a written rule set before coding began, checked September 21, 2026. A spot check of 20 coded rows against live sites, dated September 21, 2026, measured agreement between the note based coding and a direct site visit.

Score and blog cuts: median score_total by location count, 15 for one location, 17 for two, 18 for three, 20 for four or more, and blog zero share roughly 87 percent for one location versus roughly 72 percent for four or more, both checked September 21, 2026 against the same snapshot. These are associations in the current data, not demonstrated causes, and none of them should be mixed with the separate Southeast score bands from the earlier audit vintage, which used a different scale entirely.

What a Reasonable Next Step Looks Like

For an owner reading this with two or three locations and an honest sense that their own site probably falls into the overlay or missing category, the first useful step is simply checking. Open the site as a stranger would, search for the specific address of each location, and note plainly whether each one resolves to something specific or dumps the visitor onto a single generic page listing everything at once.

If the answer is a missing or broken locations page, that is the priority fix regardless of anything else in this post, since it is the pattern most directly blocking a customer from finding a specific address at all. If the answer is a single overlay page that works but crowds every location together, the next reasonable step is deciding whether a dedicated page per location is worth the effort, weighed against how different each location is from the others in hours, neighborhood, or amenities.

Not every multi location shop needs a dedicated page for each address. Two locations a few blocks apart with identical hours and nearly identical character may reasonably share a single well built page, so long as that page clearly distinguishes the two addresses and does not force a customer to guess which one is closer or currently open. The decision depends on how genuinely different the locations are, not on a fixed rule that dedicated pages are always the correct answer.

A Note on What This Study Cannot Say

This post reports counts and an honestly coded structure pattern, and it stops there deliberately. It cannot say why a given shop chose the structure it has, since the underlying notes describe what exists on a site, not the reasoning behind a decision made months or years earlier by an owner or a web developer working from constraints this data never recorded.

It also cannot say what would happen to a specific shop's score or traffic if that shop switched from one pattern to another, since the census is a snapshot of what currently exists across many different shops, not a before and after comparison of the same shop tested under two different structures. Any claim along those lines would go beyond what this data supports, and this post is not going to make it.

What it can say, plainly and with the sample sizes stated throughout, is that the location count and the coded structure pattern are both real, countable facts about the current census, snapshot dated September 15, 2026, and that a missing or broken locations page is a concrete, fixable problem independent of any question this study cannot answer about cause or effect.

Frequently Asked Questions

How many shops in the census have more than one location?

354 of 1,107 shops with a usable location count value report two or more locations, snapshot dated September 15, 2026, checked September 21, 2026. The remaining 753 report a single location. Two rows out of the full 1,109 census failed to convert to a usable number and are excluded from this count. That makes multi location shops roughly a third of the usable census.

Does having more locations correlate with a higher visibility score?

Yes, as an association in the current snapshot. Median overall score rises from 15 for one location to 17 for two, 18 for three, and 20 for four or more, checked September 21, 2026. This post does not claim adding locations causes a higher score, since longer operating history and broader investment likely explain much of the pattern.

How was page structure coded, since it is not a direct column in the data?

Among the 354 multi location shops, 333 had a gaps notes field that loosely matched language about a locations page, a per location page, or a URL pattern, coded into four bins using a written rule set before coding began: dedicated location URLs, one overlay page, locations page missing, or notes silent. A 20 row spot check against live sites measured the coding's reliability.

What is the clearest structural problem among multi location cafe sites?

A locations page that is missing or returns a broken result, since it forces a customer to abandon the site entirely and rely on Google Maps or a phone call to find a specific address. This is the pattern worth fixing first regardless of which other pattern the rest of the site otherwise follows. Checking takes minutes: open the site as a stranger and search for each address.

What is the difference between an overlay page and dedicated location pages?

An overlay page lists every location together on one page, which works but asks a customer to scan past addresses that are not relevant to them. Dedicated location pages give each address its own page with its own specific hours and neighborhood details, without competing for space with sibling locations. Two nearby shops with identical hours can still share one well built page.

Does this study prove that dedicated location pages improve search rankings?

No. No figure in Hillcane's own research supports a specific ranking claim for dedicated location pages. The association between location count and score described in this post is observed in the current snapshot, not a demonstrated cause, and this post is explicit that it makes no ranking promise. What it does say is that a missing or broken locations page is a usability gap worth fixing on its own terms.

What share of multi location shops have no blog posts at all?

Roughly 87 percent of single location shops report zero blog posts, compared with roughly 72 percent of shops with four or more locations, both checked September 21, 2026 against the same census snapshot. This is a lower blog zero share for larger multi location shops, not an absence of the problem entirely. Even at four or more locations, most shops publish nothing.

Why were four or more locations pooled into a single bracket instead of broken out further?

Pooling keeps the counts in each bracket stable enough to report honestly and avoids implying more precision than the underlying notes can support at the high end of the range. This is a deliberate simplification stated plainly, not a claim that every shop with four or more locations structures its site identically. The two, three and four plus brackets still show a steady rise in median score, 17, 18 and 20.

Can this post's numbers be combined with the earlier Southeast audit score bands?

No. This post's /40 scores come from the current census and should never be mixed with the separate Southeast score bands from the earlier audit vintage, which used a different scale entirely. Both are real, dated figures, but they describe different measurement systems and should be kept distinct. The census snapshot is dated September 15, 2026.

How reliable is the 333 row structure coding based on notes text?

A 20 row spot check against live sites, dated September 21, 2026, measured how often the note based coding agreed with a direct site visit. That spot check describes confidence in the coding method on a small sample, not a guaranteed accuracy figure for every one of the 333 coded rows, so read it as a check, not a proof.

Work with Hillcane

354 shops in the census run more than one location. Their notes show which page patterns are common, and which one is worth fixing first.

If you run more than one location and want help auditing your own site's structure, reach out at hillcane.co/contact or call (256) 384-2449.