Juliana Tavares
COO, AI&Media
Fernando Santos
Commercial, AI&Media
A 90-day structured pilot for TurfCity, Los Angeles. Google Ads and Meta Ads built around real numbers, not assumptions.
The purpose of the first three months is to answer three questions with real data. Not assumptions, not benchmarks borrowed from other businesses, not projections built on hope.
Not a click. Not a form fill. A person who wants an installation quote and has a backyard to put it in.
Artificial turf (active demand, shorter cycle) or KOP (visual discovery, longer consideration).
This number lives outside campaign management. It depends on response speed, on-site visits, and pricing. We will have the data to separate the two.
This document contains no guarantee of results. No serious media buyer can guarantee sales volume, because a decisive part of the outcome sits outside campaign management: speed of response to a lead, how the on-site visit is conducted, the price presented, and the crew's capacity to deliver. The scenarios in this proposal are projections built on public market benchmarks, with every assumption stated in the open.
Public 2026 benchmarks for home improvement and remodeling in the LA market. Every planning assumption in this proposal is derived from these figures.
Remodeling and landscaping. LA sits at the upper end due to competitive density and average job value. This proposal uses $12 as the planning reference.
Depending on market and service. Pure landscaping averages around $85. For higher-ticket services it rises accordingly.
For home services. Heavily dependent on landing page quality. This proposal plans at 5% to 10%, the lower end. The current landing pages are not yet optimized for conversion.
Per month, per platform, for a local home services business in a competitive metro area.
The split between Google and Meta is not a channel preference. It follows directly from how demand behaves for each product.
Artificial turf is actively searched for. LA homeowners look for it with purchase intent and a short decision cycle. No need to create desire, only to win the click from someone who has already decided.
KOP has limited search demand. The category is still forming in the consumer's mind. This product is won through visual discovery: show it to create desire, not compete for a search that carries little volume.
Exclusions matter as much as inclusions. Terms like diy, how to install, rolls, for sale, wholesale, jobs, cheap, home depot, soccer field attract paid clicks from people who will never hire anyone: researchers, resellers, job seekers, and amateur installers. This list goes live on day one and is reviewed weekly against actual search terms.
The management fee is fixed and not indexed to ad spend. As budget grows later, the fee stays the same and the ratio improves in your favor.
| Line item | Monthly | Quarter total |
|---|---|---|
| Ad spend, Google Ads | $1,200 | $3,600 |
| Ad spend, Meta Ads | $1,000 | $3,000 |
| Traffic management fee | $1,000 | $3,000 |
| Conversion tracking setup | One-time, month 1 only | $300 |
| Total | $3,200 (+$300 in month 1) | $9,900 |
Ad spend goes directly to Google and Meta. It does not pass through AI&Media. The management fee covers strategy, campaign architecture, optimization, and monthly reporting. It does not change as budget scales up.
Basis of calculation: $1,200 in Google divided by a $12 reference CPC yields roughly 100 clicks per month. At a 5% to 10% conversion rate, that is 5 to 10 leads per month, plus 1 to 3 qualified conversations from Meta starting in month two.
| Conservative | Optimistic | |
|---|---|---|
| Leads in the quarter | 18 | 30 |
| Assumed close rate | 20% | 30% |
| Jobs closed | ~4 | ~9 |
| Revenue generated (at $6k avg job) | ~$21,600 | ~$54,000 |
| Total investment | $9,900 | $9,900 |
Break-even: fewer than two jobs over three months.
Revenue is not margin. At a 40% gross margin, a $6,000 job leaves $2,400. Real break-even moves to roughly four jobs in the quarter. Recalculate with your actual numbers.
Close rate is not controlled by traffic management. It depends on response speed and how the on-site visit is handled. If leads arrive as projected and jobs do not close, the diagnosis will not be in the traffic, and we will have the data to demonstrate that.
We estimate 6 to 10 assets per month to feed the tests. Part of this can be repurposed from content already produced for the profile. Ad creative is not feed creative.
3D renders do not convert on high-ticket products. A Beverly Hills buyer wants to see the object installed in a backyard, not a simulation. At least half the creative mix must be real footage.
The product must be seen in context. An outdoor kitchen without people in it does not communicate the lifestyle. This is what Meta is buying.
Transformation content consistently outperforms product showcase in home services. Required each month.
Assets must be delivered no later than five business days before each monthly cycle begins. We reserve the right to request replacement of assets that do not meet technical specification for format, duration, or aspect ratio.
This is below market reference for work of this complexity, because this is a pilot project. It does not recur in months 2 or 3.
Google Tag, Google Ads conversions, Meta Pixel, and Meta Conversions API. End-to-end attribution from click to conversion.
The dominant conversion channel in this business. Without it, phone calls are invisible to the algorithm.
Pipeline stages fed back into Google and Meta through offline conversion events. Without this, the algorithms optimize for form fills and deliver low-quality volume. With it, they optimize for closed estimates. That is the difference between spending budget and investing it.
Outside the contracted scope. Raised because they directly affect what you pay per click, and whoever manages the website should know about them.
The service area pages share identical content with only the city name swapped. This degrades Google Ads Quality Score. A low Quality Score means paying more per click than competitors pay for the exact same ad position.
Website, Google Business profile, and Yelp show different information. This hurts local organic search and the performance of location extensions in Google Ads.
A $19,900 product currently has no identity of its own in the eyes of search engines or AI assistants, which are increasingly where high-ticket purchase research begins.
This is an alert, not an audit. A structured SEO, GEO and AEO review can be proposed as a separate phase after the first traffic cycle, when there will be real search data to ground it.
These are not a formality. The answers directly determine how the campaigns are structured and how performance is measured.
Clearly defining scope protects both sides. What is contracted, what sits outside, and what will never be promised.
Guaranteed sales volume, guaranteed number of leads, or guaranteed cost per lead. The figures in this proposal are projections based on public market benchmarks with stated assumptions. They are subject to validation against real data from month two onward.
A decision made before the end of month two would be made without enough data to be correct. The pilot is designed to generate the evidence, not to demonstrate results prematurely.
| Period | Focus | What gets evaluated |
|---|---|---|
| Month 1 | Setup and signal collection | Conversion tracking configuration, GoHighLevel integration, search term cleanup, negative keyword list. Not evaluable for campaign performance: this is the algorithmic learning phase. |
| Month 2 | First real read | Cost per lead by platform and by keyword cluster. First structural optimizations. Initial KOP prospecting data. |
| Month 3 | Decision | Cost per closed job. Full quarter recommendation: scale, reallocate between platforms, or stop. |
Ready to move forward? The next step is confirming the answers to the questions in section 9 and scheduling a setup call.
Let's get started →