The same case study as above, in plain text for screen readers, search engines and anyone who prefers to skim.
Case study / Boston, Massachusetts
MAX Ultimate Food is one of Boston’s most established catering companies. Across two years its marketing program was rebuilt from the website up: three Best of Boston wins amplified, paid social cut by roughly 80 percent, and every lead in the business put under an AI-assisted microscope. Volume held. Quality moved.
Photography courtesy of MAX Ultimate Food
3
Best of Boston awards won and amplified across twelve months
+66%
Website-attributed events created, January to June, year over year
94%
Of web leads arriving on organic or direct URLs
−80%
Paid social spend, with total lead volume holding steady
01 The brief
MAX Ultimate Food caters the city’s landmark rooms, holding named venue relationships at institutions including the Boston Children’s Museum, the Museum of Science, and WGBH. Its reputation with planners is excellent. Its close rate is exceptional.
A full read of the sales year showed a business winning 52.8 percent of the deals it worked, against an industry norm nearer 20 to 35 percent, and turning 600 confirmed events into roughly 8.6 million dollars of revenue. The problem sat on either side of the kitchen. Upstream, paid social was buying enormous volumes of leads that almost never became customers. Downstream, 537 cancellations took 12.4 million dollars of pipeline with them, and roughly three quarters of those died before any real engagement.
One form told the whole story. The Book a Tasting form on the old website generated 117 leads in and produced zero qualified opportunities and zero confirmed events.
The product was never in question
Volume was not the constraint. Everything downstream of volume was.
02 Scope of work
The engagement ran as a single integrated retainer rather than a set of channel projects. Research set the priorities, the website carried the conversion load, earned media supplied the demand, and the analytics layer decided every quarter where the money went next.
Research & analysis
Salesforce trackers, the INFOR booking system, five website forms and GA4 exports reconciled into one view of every inquiry, event and dollar.
1,046 leads and 600 events analyzed for the sales year
Website
A WordPress rebuild with a custom venue content type, event-type to venue mapping, and five purpose-split lead forms replacing one catch-all.
47 venue pages live at maxultimatefood.com
Search engine optimization
Technical audit, keyword research, title and meta rewrites, internal linking, and an editorial calendar aimed at planner intent rather than vanity terms.
Organic search grew from 15% to 34% of sessions
AEO & GEO
AI crawlers were being refused at the server. Once unblocked, the site was restructured so assistants could read and cite it.
FAQ schema doubled to 31 pages, 148 questions
Paid digital
Meta lead ads were producing volume without qualification. Spend was reduced sharply and pushed toward high-intent search and award-window retargeting.
Paid social fell from 63% to 21% of sessions
Public relations
Three Best of Boston wins treated as planned campaign windows rather than press releases, with amplification timed to the print issue and social echo.
Wedding inquiries up 92% after the Weddings win
Organic social
Monthly on-site photo and video production feeding a three to five post weekly cadence plus a weekly reel across Instagram and Facebook.
51% engagement vs paid social’s 30%
Graphic design
Print advertising across eleven trim sizes, catering and wedding guides, venue one-pagers, sales kits, menus, postcards and the MAX Out sub-brand.
Roughly 30 design programs across 20 months
Automation & sales
Speed-to-lead SLAs, deduplication, UTM capture, form to CRM sync and nurture sequencing specified from the loss data rather than from best practice.
67% of all losses traced to non-response
03 The channel flip
In the paid-social year Meta lead ads delivered 2,309 leads at $18.74 each, which looks like a bargain until you follow them. Seventy-five became qualified opportunities. That is $577 per qualified lead.
In the year after, that spend footprint was cut by roughly 80 percent. Sessions barely moved on a run-rate basis and lead volume held, because direct, organic search, referral and earned media stepped into the space paid social vacated. Together the two paid channels now account for under 6 percent of form leads.
Share of website sessions by channel. Year one is the full paid-social year; year two covers January 1 to August 6.
Year 1 Year 2
Source: GA4 channel exports
Total advertising spend across Meta and Google, trailing twelve months.
Source: MAX Ultimate performance table, Meta and Google Ads
$577
True cost per qualified lead from Meta, against a headline cost per lead of $18.74
2.4%
Session to inquiry rate in the year after, strongest in the months where organic share was highest
65%
Paid search engagement rate, up from 30 percent a year earlier on a small, efficient budget
04 Earned media as a channel
Boston magazine named MAX a Best of Boston winner three times in twelve months. Most agencies would issue a release and move on. Each win was instead instrumented, measured against equal windows before and after the announcement, and treated as a planned campaign moment.
The finding that changed the calendar: the lift does not peak on announcement day. It peaks three to four weeks later, when the print issue circulates and the social echo lands. Late January and late July are now built as amplification windows, with award badging on the site and in proposal templates, and email to the house list timed to the echo rather than the press release.
The first win had originally looked like a dud, because the first analysis used booking-system records that lag inquiries by weeks. Going back to the raw form exports found a 47 percent lift that the booking data could not see.
Award creative, produced in house
First win, announced
+47%
Website form submissions across eight-week windows, 66 to 97. Every post-award entry arrived organic or direct, before any paid campaign existed.
Weddings win, published
+92%
Wedding inquiries in the four weeks after versus before, 12 to 23. Independently confirmed by raw form entries, which rose 110 percent.
Repeat win, announced
+92%
Daily new users during the surge, 70 per day before the announcement to 134 per day after it. Tracker inquiries rose 20 percent to a record July.
Logged inquiries from the master tracker. August is excluded because the reporting period covered only six days. Award announcements marked.
Source: master email and phone inquiry tracker, January 1 to August 6
July’s 78 inquiries is the best month in the tracker’s history, and no paid campaign sat behind it. The previous peak, 79 in, was bought.
05 Answer engine optimization
Planners increasingly ask an assistant before they ask a caterer. The audit found the site was returning a 403 to AI crawlers at the server level. It was, to that entire class of visitor, invisible.
Once access was restored, the site was rebuilt as a source worth citing. Structured FAQ data went from 15 pages carrying 77 questions to 31 pages carrying 148. Seven pages had visible FAQs that crawlers could not read at all. Nine main pages had none.
The single most common question people ask an assistant about a caterer is what it costs. That answer lived on four low-traffic pages. It now appears on seven. The company entity was rewritten to carry both Boston magazine awards, a six-state service area, twelve services and NACE membership, and a plain-text summary for assistants was published at /llms.txt.
148
Structured questions across 31 pages, up from 77 across 15
71%
Engagement rate from AI assistant referrals, the highest of any channel on the site
18 ms
Time to first byte, with full page load under one second
Accessibility and machine-readability work completed. Contrast figures are measured ratios against WCAG minimums of 4.5 to 1 for body text and 3 to 1 for large headings.
Item
Before
After
Result
AI crawler access
403 refused
Open
Fixed
Pages carrying FAQ structured data
15
31
+107%
Structured questions answered
77
148
+92%
Pages answering catering cost questions
4
7
+75%
Header contrast ratio
1.83: 1
7.67: 1
Passes AA
Headline contrast ratio
1.49: 1
4.82: 1
Passes AA
Venue pages with a working skip link
0 of 47
47 of 47
Fixed
Photos with real alt descriptions
Filenames
28 rewritten
Fixed
Duplicate meta descriptions per blog post
2
1
Fixed at source
Plain-text assistant summary
None
/llms.txt
Published
Source: technical and AI-visibility remediation log
06 The growth segment
Weddings became the fastest-growing part of the business: 32 percent of all typed inquiries, with wedding share of typed form submissions rising from 26 percent to 37 percent year over year.
They are also the easiest segment to misjudge. Only 4 percent of wedding inquiries have booked, with 45 still open, because a wedding inquiry raised this year books a date one to two years out. Judged on same-year booking rate, the best-performing segment in the business looks like the worst.
The right scorecard is pipeline created, and on that measure wedding-occasion event records rose 52 percent. A wedding lead magnet launched on July 23 captured 18 consideration-stage leads in its first 48 hours. The Knot, kept funded on evidence rather than habit, accounts for 11 booked-event referrals.
Wedding service
Event records created in the booking system. Both measures count events created, so they share a single scale. A July 21 bulk data-entry event is excluded from every figure.
Year 1 Year 2
Source: INFOR booking reports. Wedding occasions January to July, website referrals January to June
07 Craft
Around thirty distinct design programs ran across twenty months: print advertising built to eleven separate trim sizes, a catering guide taken through nine rounds, a wedding catering guide, venue one-pagers, menus, sales kits, postcards, newsletter templates and the MAX Out sub-brand. Monthly on-site photo and video production kept the social channels fed with real events rather than stock.
Organic social returned a 51 percent engagement rate against paid social’s 30 percent, at no media cost at all.
Two new exclusive venues launched inside the existing site rather than as standalone properties
08 The finding
Of 167 lost inquiries, 112 were lost as unresponsive. Two thirds of everything the marketing program generated died because nobody got back to it in time.
That is not a media problem and no amount of additional traffic fixes it. The recommendation that came out of the analysis was deliberately unglamorous: a same-day first-touch commitment, a three-touch automated follow-up on every web inquiry, real-time deduplication on email and phone, and a direct form to CRM sync so that the 31 percent of inquiries that never reached the CRM at all stop disappearing.
Recovering even a fifth of those 112 unresponsive losses would out-produce both paid channels combined. It is the cheapest lead-volume increase available to the business, and it costs no media at all.
67%
Of all lost inquiries closed as unresponsive rather than lost to a competitor
52.8%
Win rate on deals actually worked, against an industry norm of 20 to 35 percent
$4–5.75 M
Annual revenue upside quantified from closing the response and qualification gaps
09 Results ledger
Measure
Prior period
Current
Change
Website-attributed events created, January to June
29
48
+66%
Wedding-occasion events created, January to July
75
114
+52%
Wedding share of typed form submissions
26%
37%
+11 pts
Organic search share of sessions
15%
34%
+19 pts
Direct share of sessions
16%
35%
+19 pts
Paid social share of sessions
63%
21%
−42 pts
Paid search engagement rate
30%
65%
+35 pts
Events reaching booked stage, January to June
335
333
Held
Website form submissions, January to early August
365
318
−13%
Total tracked leads, January to early August
n/a
479
461 tracker + 18 magnet
Raw form volume fell 13 percent, and that is the point. The prior year’s total included a large body of generic, untyped, paid-social traffic: 66 percent of entries did not even specify an event type. The cohort self-identifies, skews toward weddings, and converts into materially more booked business.
Method
Figures are drawn from the client’s own systems: the master email and phone inquiry tracker, five salesperson Salesforce trackers, the INFOR booking system, Gravity Forms exports across five website forms with spam removed, GA4 acquisition and engagement exports, Meta and Google Ads reporting, and Boston magazine award announcement dates. Year-over-year traffic comparisons are not drawn, because GA4 collection on the property began August 8,. A July 21, bulk data-entry event in the booking system is excluded from every trend. Reported GA4 users carry roughly 19 percent data-center inflation, which affects traffic figures and not inquiry counts.
Period
Website rebuild begun at the outset. Current marketing services agreement effective September 15,. Performance analysis covers calendar in full and January 1 through August 6,.
Credits
All photography courtesy of MAX Ultimate Food. Confidential. Prepared for discussion.