Case Study · Montgomery Water Softening · Residential water treatment, Conroe TX · Engagement signed 2026-04-20, ongoing
Week 1: zero ad spend, and every core engagement metric multiplied – up to 8.8x.
Lakeside Lender signed Montgomery Water Softening on 2026-04-20. Google Ads did not serve a single paid impression until 2026-05-08 – so everything in this table was earned in the first engagement week from foundation work alone: on-page SEO, structured data, city and service landing pages, an AI assistant, and Google Business Profile work. Figures verified against the client’s GA4 property.
| Metric | Pre-engagement (2026-04-14 to 04-20) | Week 1 (2026-04-21 to 04-27) | Change |
|---|---|---|---|
| Active users | 15 | 72 | +380% (4.8x) |
| Page views | 46 | 407 | +785% (8.8x) |
| Conversion-grade events | 3 | 22 | +633% (7.3x) |
That was the foundation week. Here is what it compounded into.
The starting point
A family-run Conroe-area water treatment company whose owner wanted growth. The site was published but underbuilt: bleeding SEO equity after a platform migration, no structured data, a homepage still headlined under the previous company name, city links that all dumped into one contact form, and a stale, misconfigured Google Ads account. The brief: fix the foundation, build the engagement layer, and launch paid cleanly – without breaking what already worked.
Where the account stands today
Every figure carries the date it was read.
| Signal | Reading | As of |
|---|---|---|
| Google Ads, trailing 30 days | 7,609 impressions · 426 clicks | 2026-08-18 |
| Impression share lost to rank | 0.00% | 2026-08-10 |
| GA4, trailing 28 days | 742 active users · 934 sessions | 2026-08-18 |
| Google Business Profile | 70 reviews · 4.9 stars · 664 customer interactions (trailing 6 months) | 2026-08-18 |
| Monty, the AI assistant built for the site | 24 multi-turn customer conversations logged | 2026-08-07 |
| New /careers hiring page | Published, application path verified end to end | 2026-08-18 |
Eight pieces of work that moved this account
1. From zero paid impressions to 7,609
The account had never served a paid impression – a launch-blocking hold sitting above the campaign layer, the kind of problem most agencies never even find. Lakeside escalated it, and the first paid impressions in the account’s history landed on 2026-05-08, eight days later. That same account now runs 7,609 impressions and 426 clicks per trailing 30 days (2026-08-18).
2. A bidding doom loop, diagnosed and reversed
Both campaigns had been pushed onto conversion-based bidding without the volume to train it. Lakeside diagnosed the loop and reverted them to Maximize Clicks – and in the following week, cost per conversion fell 56% ($81.08 to $35.57) while conversions rose from 4 to 10 and search impression share jumped from 10.5% to 26.5%. Both sides of the ratio moved the right way in the same week.
3. Site traffic more than doubled in four weeks
The client’s own site analytics tracked the ramp independently: rolling 7-day visitors climbed from 493 to 1,032 and rolling 30-day visitors from 1,010 to 2,242 between 2026-04-27 and 2026-05-21.
4. Rebuilt for search: indexed pages, 5 to 26
The site went from nearly invisible to crawlers to fully structured: sitewide LocalBusiness and FAQ schema, meta descriptions on every page, a rebuilt homepage, thirteen city and service landing pages, and an eight-phase technical deploy. Indexed pages grew from 5 to 26 and indexed city and service pages from 1 to 17 (2026-05-06 to 06-12), with structured data brought to passing validation – the eligibility bar for star ratings on the results page.
5. Google Ads optimization score: 99.6%
Google’s own account-level optimization score went from 70.3% to 99.6% in a single working session (2026-05-18), with both campaigns reading above 99% two days later. The auction agrees: staged budget and schedule moves drove impression share lost to rank to 0.00% (2026-08-10). The account stopped losing auctions on quality – the only remaining constraint is budget, which is a far better problem to own.
6. The brand owns its own search results
From barely indexed to #1 with sitelinks on the brand name (confirmed 2026-06-18), with 8,350 search impressions in the trailing 28 days in Search Console. When a competitor claimed otherwise to the owner, the live check settled it.
7. Local presence: 70 reviews at 4.9 stars
Google Business Profile reviews grew from 51 to 70 with the 4.9-star average held throughout, and customer interactions on the profile more than doubled – 271 to 664 on a trailing six-month basis (2026-05-04 to 2026-08-18).
8. Built two moves ahead of demand
Growth capacity was built before it was needed. A /careers hiring page shipped 2026-08-18 with the application path verified end to end – no hiring path had existed on the site before it. And a complete expansion package for the Tomball-Cypress corridor – nearly 100,000 target households, six budget scenarios, thirteen ready-to-import campaign files – is built, independently verified, and staged for the owner’s green light.
How the work is governed
Every change above shipped under the same discipline: one experiment at a time, a written theory before the change, a pre-declared abort condition, and guardrails checked the next morning. The full standard – change discipline, call gates, automation guardrails – is documented on the AI Stack page.
The day-by-day build record
The full first-week operational story – the exact slug list for the city and service pages, the conversion-action definitions, the ad-group structure, the AI assistant’s system-prompt design, and the timing call on when to enable paid spend – lives in the gated deployment playbook, for prospects evaluating a Dominate-tier engagement.
Request the MWS deployment playbook ->
Want the same build-out for your business?
30-minute discovery call, then a free audit, then a written recommendation. No follow-up sequence, no pressure.