Case Study · Roxbury, Boston
Google showed them less often than before. More people clicked anyway.
Nubian Markets is a halal market and café on Washington Street in Roxbury. Between June 28 and July 25, 2026, Google showed their pages 28% less often per day than in the weeks before, and clicks per day still went up. The share of people who clicked rose from 3.4% to 4.9%. Honey Bridge changed what those people saw before they clicked.
The result
Shown less. Clicked more.
Google showed Nubian Markets 568 times a day, down from 789. Clicks went from 26.7 a day to 27.8.
Search Console counts two separate things. How many times your pages appeared in front of somebody, and how many of those people clicked. Between June 28 and July 25 the first number fell 28% per day. The second one still rose. Both bars are per day and indexed against June 8 to 27, so the dotted line is the weeks before.
When fewer people see you and more of them click, the thing that changed is what they saw. That is the sentence Google prints under your name, and it is the part most restaurants never write.
Google Search Console for nubianmarkets.com · Jun 28 to Jul 25, 2026 against Jun 8 to Jun 27, 2026, per day
| What Google counted | Jun 8 to 27 | Jun 28 to Jul 25 | Change |
|---|---|---|---|
| Share of the people shown who clickedClick-through rate | 3.38% | 4.89% | +45% |
| People who clicked through, per dayClicks | 26.7 | 27.8 | +4% |
| Times Google put them in front of someone, per dayImpressions | 789 | 568 | −28% |
| Average spot in the resultsLower is better | 8.4 | 8.9 | Worse |
Measured 2026-07-27
Source: Google Search Console, property sc-domain:nubianmarkets.com, for Nubian Markets in Roxbury, Boston. Jun 28 to Jul 25, 2026 compared against Jun 8 to Jun 27, 2026, the first 20 days the property holds, per day. Raw totals: 533 clicks and 15,776 impressions over 20 days, then 778 clicks and 15,897 impressions over 28 days. Pulled 2026-07-27, corrected 2026-10-05 by Honey Bridge.
What we found
Three faults on the Nubian Markets site. You can check all three on yours in a few minutes.
The address on their own site was wrong.
Their About page printed 2656 Washington St. The real address, the one on their Google listing, is 2565. A business listing and its website disagreeing on the address is one of the first things a local SEO audit checks.
Check yoursRead the address on your website out loud against the one on your Google listing. Digit by digit.
The homepage was serving 12 placeholder images.
A broken Instagram block sat on the homepage loading a dozen stock placeholders instead of their food. Grey boxes and filler captions, above the fold, on the first page anyone lands on.
Check yoursOpen your homepage on your phone on cell data, not wifi. Anything that does not load is doing this to you.
10 of their 18 pages had no description written for them.
The description is the grey sentence Google prints under your name in the results. Nobody had written one for most of the site, so Google assembled its own out of whatever text it found first. The About page had none at all.
Check yoursGoogle your restaurant name. Read the sentence under the link. If it reads like a fragment cut off mid-thought, nobody wrote it.
What we changed
Six changes to the Nubian Markets site on July 7. Then more on the 22nd and the 24th.
- Corrected the About page address and its map pin to 2565 Washington St, matching their Google listing.
- Wrote the titles Google prints for the homepage and the About page.
- Wrote the About page description. It had been empty.
- Added structured data across the site stating, in the format Google reads, that this is one halal grocery store and restaurant at one address in Boston.
- Added a real headline to the homepage, the About page, and the Nuba Café page.
- Deleted the broken Instagram block and its 12 placeholder images.
July 22, 2026
- Added Butcher shop and Event venue to their Google Business Profile categories, which is how halal meat and private event searches reach a listing at all.
July 24, 2026
- Descriptions written for 17 of 18 pages, up from 8.
- Alt text on all 52 images, up from 31.
- Removed an internal page from the search index.
- Published their first blog post under a real author byline.
The 28-day window covers 19 days of the July 7 pass, 4 days of July 22, and 2 days of July 24. Whatever the last two are worth, almost none of it is in the number at the top of this page. We would rather say that than let you find it.
That order is not arbitrary, and it is the order we would run on your restaurant. The listing and the address first, because they cost nothing and Google weights them heavily. Then the words it prints under your name. Then the structured data that tells it what you actually are. Then the pages.
What the numbers rule out
Exposure fell. Clicks did not. That is the part that rules things out.
A good month for halal food in Boston would have pushed impressions up. They went down.
When a season, a holiday, or a neighbourhood gets busier, Google shows you to more people. Impressions rise. That is what a rising tide looks like in this report, and it is not what happened here. Google showed Nubian Markets 789 times a day, then 568. The tide did not come in. What changed is the share of people who chose to click, 3.38% to 4.89%.
More demand moves impressions. A better snippet moves the click rate. Only the second one moved. That is the whole argument, and it is made entirely out of Nubian Markets’ own report. The split below is the same argument run a second way.
The obvious objection: were those just people typing “nubian markets” into Google?
Somebody who already knows the name and searches for it is not a win we can claim. So we pulled the report again with every search containing their name filtered out, and looked at what was left.
| Segment | Clicks | Times shown | Share who clicked | Average spot |
|---|---|---|---|---|
| Every searchMeasured | 26.7 → 27.8 a day, +4% | 789 → 568 a day, −28% | 3.38% → 4.89% | 8.4 → 8.9 |
| Searches with no mention of their nameMeasured | 1.25 → 1.89 a day, +51% | 167 → 188 a day, +13% | 0.75% → 1.01% | 14.9 → 14.3 |
| Searches that used their nameDerived by subtraction | 25.4 → 25.9 a day, +2% | 622 → 380 a day, −39% | 4.08% → 6.81% | Not available |
Clicks per day from searches with no mention of their name rose about half, 1.25 to 1.89, and their average spot in those results improved. On searches that did use the name, Google showed them 39% less per day and they still got slightly more clicks. Fewer chances, more clicks, is what rewriting a title and a description looks like in a report.
Search Console reports filtered totals as approximate. The no-name row is measured directly with a query filter. The used-the-name row is derived by subtraction and may be slightly off.
What this page does not prove
Five things a careful reader should hold against these numbers. We would rather print them than have you find them.
Their average spot in Google’s results got slightly worse over the window, 8.4 to 8.9. Clicks went up anyway, because more people picked them from the same place on the page. We did not move them up. We made them worth choosing.
Most of these searches were people who already knew Nubian Markets by name. 725 of the 778 clicks. We had not yet proven we could bring you people who have never heard of you. By October 2026 we had: the follow-up case study.
This is one 28-day window. Twenty-eight days is long enough to see a change and too short to call it permanent. The date we re-check is printed below.
Every number in the result above, and in the split below it, comes from one property. We are not showing you a comparison against other restaurants over the same window, so you are taking our word that nothing else changed. The argument rests on the shape of Nubian Markets’ own report, not on a control group.
The searches with no name in them start from a small base. 53 clicks over 28 days is about two a day. The percentage is large because the number underneath it is small, and a percentage on a small number is the oldest trick in this industry.
The next number
Anyone can show you one good month. Here are the two we committed to in advance, and the one we have kept so far.
Promise one · 2026-10-25
We will pull this exact report again on October 25, 2026.
Same property, same 28-day comparison, same four rows. A lift that holds for three months is a different claim from a lift that holds for four weeks, and only one of them is worth paying for. Whatever the October report says goes on this page, in this spot.
Promise two · kept 2026-10-05
Fries B4 Guys, 90 days after their new site: more people found them, and a smaller share of them clicked.
Fries B4 Guys moved from WordPress to a new site on July 20, 2026. Their old site had pages with no title, no description, and no headline, and Google had indexed 8 of its 44 pages. We printed the baseline on this page before we knew the result, and promised to publish September against it, including a decline. One correction to our own baseline first: Search Console only holds Fries B4 Guys data from April 28, 2026, so the 90 days we printed in July really held 82. We compare per day below so both windows are measured the same way.
| What Google counted | Apr 28 to Jul 18 | Jul 3 to Sep 30 |
|---|---|---|
| People who clicked through, per day | 24 | 25, up 4% |
| Times Google showed them, per day | 398 | 646, up 62% |
| Share of those people who clicked | 6% | 3.9%, down |
| Average spot in the results | 6.9 | 7.3, worse |
| Pages Google had indexed | 8 of 44 | 9 of 10 |
The decline. The share of people who clicked fell from 6% to 3.9%, and their average spot slipped from 6.9 to 7.3. Daily clicks barely moved. Most of the drop is on searches for their own name: Google showed those 53% more often per day, and they were clicked 2% less. We do not yet know why, and we are not going to guess on this page. On searches that did not use their name, clicks per day rose 53%, from about 2.7 to 4.1.
Source: Google Search Console, property https://friesb4guys.boston/, for Fries B4 Guys in Allston, Boston. The baseline covers April 28 to July 18, 2026, the full days the property holds before the move (82 days). The new window covers July 3 to September 30, 2026 (90 days), measured 2026-10-05 by Honey Bridge; it includes the 17 days before the July 20 move. Pages indexed is 8 of 44 on the old site against 9 of the 10 pages the new site lists for Google, checked with the URL Inspection API on 2026-10-05.
Work with us
Want that sequence run on your restaurant?
Listing and address, then the words Google prints under your name, then the structured data, then the pages. None of it is complicated. Start by seeing where you stand today.
Call for a free health check