TeamWhat We DoCase StudiesToolsBook a consultation →
Restaurant interior — hospitality context
CASE STUDY. TOCK. HOSPITALITY · B2B SAAS

Less Spend.More Demos.

Tock sells reservation and experience software to restaurants and hospitality operators. iExcel rebuilt the measurement spine, then the paid engine — and cut cost-per-demo 63% in one month.

Engagement
Jul 2023 — Mar 2024
Phases
HubSpot foundation · Paid + SEO scale
Channels
Google · Bing · Meta · LinkedIn · YouTube · SEO
Stack
HubSpot · Salesforce · GTM · GA4
// The scoreboard · what the engine delivered
−63%
Cost per demo · one month
The December reset: budget pulled back, search account restructured around what converts.
+54%
Demo requests · Jan → Feb
While cost-per-demo fell another 4.6% against the reset baseline.
−70%
Cost per demo · Dec → Mar
Three consecutive monthly declines against the program’s own baseline.
21
First-page rankings · Mar 2024
From 69 newly optimized resource pages — 27 more queries queued on page two.
The Stack

Tools we ran it on.

HubSpot
CRM · Workflows · Lead scoring
Salesforce
Sales handoff · Attribution
Google Ads
Search · Broad-match testing
Microsoft Ads
Bing search
Meta Ads
Paid social · Lookalikes
LinkedIn Ads
B2B reach · Insight Tag
YouTube
Video action campaigns
GTM / GA4
Conversion tracking · QA
Brizo + Apollo
Audience enrichment · ABM
Chapter 01 · The Setup

A modern restaurant runs on software.

Tock's buyer isn't a casual diner. It's a sommelier, a GM, a hospitality group COO. Sales cycles are deliberate. Onboarding requires trust. A demo request is the first real signal of intent — and every one of them has to be earned.

When iExcel came in, the surface looked busy. Google Ads were running. LinkedIn was running. Forms were submitting. But underneath — the measurement spine that ties campaigns to demos to closed customers wasn't there yet. You can't optimize what you can't trust.

So we did it in two phases. First, build the foundation. Then, turn the engine on.

Restaurant dining room
// Hospitality SaaS · Reservations · Experiences · Wineries · Fine dining
Chapter 02 · The Two Phases

Foundation first. Engine second.

Most agencies start with the ad account. We started one layer deeper — with HubSpot, Salesforce, and the question every marketing leader needs answered before they spend another dollar: which demos are real, and where did they come from?

Phase 01 · Foundation
JUL 2023 — FALL 2023

HubSpot Marketing Cloud Enterprise.

The implementation everything else depends on. Lifecycle stages defined. Sales handoff mapped. Attribution wired.

  • Lifecycle, lead-status, and attribution definitions
  • Salesforce field mapping + integration scope
  • Legacy email-platform migration + enrichment plan
  • HubSpot forms, workflows, smart lists, lead scoring
  • GTM-based HubSpot pixel + UTM governance
  • Channel integrations: Google · Meta · LinkedIn · Bing
Phase 02 · Engine
NOV 2023 — MAR 2024

Paid · SEO · Conversion Tracking.

With the spine in place, we turned the channels. Audit, restructure, test, scale. Then verify every conversion was real.

  • Google Search audit + restructure + broad-match testing
  • Bing, Meta, LinkedIn audits + audience rebuilds
  • Brizo + Apollo audience enrichment for ABM
  • SEO: 105 pages keyword-mapped, 69 pages optimized
  • 112 pages categorized for schema markup
  • GTM / GA4 / LinkedIn Insight Tag QA + repair
Restaurant bar seen through its street window at night, with RELAX lettering on the glass
// The hardest part of B2B hospitality marketing

"The buyer is busy running a restaurant. Your one shot is a clean, fast, trustworthy demo request — and a measurement layer that proves it converted."

Chapter 03 · The Reset

Same audience. Half the spend.

December 2023 was the inefficiency peak. Cost-per-demo on Google Search had drifted to the worst month on the program's books — the kind of number a paid account earns when it's spending against everything instead of what converts. So we pulled the budget back, restructured the search account around the campaigns that were actually producing demos, and shipped a tighter spend plan into January.

The result was the kind of month every paid lead asks for and rarely sees: against the program's own December baseline, spend fell 56%, demo requests rose 18%, and cost-per-demo dropped 63% — in a single monthly cycle. Not by cutting the volume. By cutting the waste.

Google Search · Cost per Demo Request, indexed
// DEC 2023 = 100 · BAR LENGTH = RELATIVE COST PER DEMO
DEC '23
100
BaselinePre-reset
JAN '24
37
−63% vs DecThe reset
FEB '24
35
−65% vs DecThe scale
MAR '24
30
−70% vs DecThe refinement
// THREE-MONTH STORY · RESET → SCALE → REFINE · INDEXED TO THE PROGRAM'S OWN DECEMBER BASELINE
+18%
Demo requests · Dec → Jan
Delivered while paid spend was cut by more than half against the program's own December baseline. The waste left; the volume stayed.
Chapter 04 · The Scale

Then we pressed harder.

The reset proved the search account could be efficient. February proved it could scale without sacrificing the efficiency. Spend came back up 47% month-over-month, demo requests rose 54% — and cost-per-demo still fell another 4.6%. Spend up and cost-per-lead down in the same month is the combination most paid programs never print; it only happens when the account structure is doing the work, not the budget.

The biggest unlock was the competitor campaign. In January it had been the least efficient line in the entire account. In February — restructured, retargeted, refined — its cost-per-demo came down roughly 85% against its own January baseline, from the worst unit economics in the test matrix to among the best.

Broad-match testing on the non-brand reservation-software campaign hit at the same time: demo requests up 124% at a 10% lower cost-per-demo. The kind of result that doesn't come from "more spend." It comes from a better account.

+54%
Demo requests · Jan → Feb 2024
While cost-per-demo fell another 4.6% against the January reset baseline — growth that didn't buy itself with efficiency.
Busy upscale dining room seen from above, every table in service
// February · Scale without sacrificing efficiency
Chapter 05 · The Refinement

The line kept bending.

March pulled cost-per-demo down another 13.2% — the third consecutive monthly decline, stacking to roughly 70% below the December baseline — while demo volume held near its February peak. The US non-brand competitor line kept improving too: demo volume more than the prior month, at a lower cost-per-demo.

Paid social read cleanly against the same yardstick. Meta was the early efficiency winner — the lowest cost-per-conversion of any paid channel in January, with a 7.6% form-view-to-submission rate on the Experience Playbook campaign. LinkedIn drove reach at seven-figure monthly impression volume but needed budget consolidation to optimize properly. We flagged it, structured the recommendation, and handed it off.

−70%
Cost per demo · Dec 2023 → Mar 2024
Three consecutive monthly declines, measured against the program's own December baseline. The reset held — then kept compounding.
Modern restaurant dining room with warm copper tones, set tables, and an open kitchen line
// March · The third consecutive monthly CPA decline
// Cost per Demo Request
63%

Down in a single month.

December 2023 to January 2024. Spend cut by more than half. Demo requests up 18%. Cost-per-demo down 63% against the program's own baseline. The kind of step-change that doesn't survive without a measurement layer to prove it's real — and that's what Phase 1 built.

Chapter 06 · The Compounding Half

105 pages, mapped to a plan.

Paid is the loud half. SEO is the compounding half — and on a B2B SaaS site selling into restaurant operators, the compounding has to be earned page by page.

We mapped 105 pages to a keyword strategy, rewrote metadata and H1s across 69 resource pages, categorized 112 pages for schema markup, and shipped a hub-and-spoke plan plus a keyword-gap roadmap. By March, the Resources section was holding 21 first-page rankings, with 27 more queries sitting on page two — the clearest kind of next-quarter growth runway an organic program can show.

Rankings landed on the queries that matter to operators: "5 steps to launching POS," "Opening Restaurant 90 Day Checklist," "importance of restaurant POS." Homepage caught permutations of "restaurant reservation system."

21
First-page rankings · Resources section, Mar 2024
Built from 69 newly optimized resource pages — with 27 more queries on page two queued up as the next-quarter runway.
Independent cafe interior with a service counter, menu boards, and guests at wooden tables
// SEO foundation · 105 mapped · 69 optimized · 112 schema'd
Chapter 07 · The Measurement Foundation

The conversions we didn't count.

Every paid number in this case study survived a QA pass first. Some of them only exist because of one.

The LinkedIn Insight Tag's conversion ID was firing on pageview, not action — every page load was being counted as a B2B form submit. We caught it in GTM, isolated the misattached ID, and corrected the tracking. Cleaner numbers. Defensible reporting. A conversion count that finally matched a real lead-per-user ratio.

Same diligence on the GA4 side. Form-submit conversions had quietly fallen off after January 11 when the site moved to HubSpot forms — GTM triggers were still pointing at the old form structure. Fixed. Documented. Flagged for the team.

This is the work that doesn't show up in a dashboard. It's the work that makes the dashboard worth reading.

Two analysts working side by side at code and dashboard screens in a bright office
// GTM · GA4 · LinkedIn Insight Tag · HubSpot · Salesforce
Chapter 08 · The Read

Why this story matters.

For B2B SaaS leaders selling into operators with no patience for noise — restaurants, hotels, regulated industries — the Tock engagement is the template.

It's the story of foundation before flash. CRM before campaigns. Attribution before optimization. Conversion-tracking QA before any number gets reported. Then — only then — the paid engine, tuned hard enough to drop cost-per-demo 63% in a single month and scale it from there.

Boutique team. Senior operators. Two phases, nine months, one engine.

Foundation first.
Engine second.

iExcel runs the CRM implementation, the attribution layer, the paid engine, and the QA spine that makes every reported number defensible.

Book a 15-min sanity check →
Book →