A metal-art client was selling custom family-name pieces while copying ads built around a competitor's founder. Our account data gave us a different direction. I exported the Facebook ad records, compared buyer groups in Google Sheets, and used those findings to change the message and the campaign structure. The work helped take the store from around $10K a month to around $150K a month in about 90 days.
A customer could put their family name on a piece of metal art and hang it at home. But the brand was barely breaking even, and a successful competitor had become the reference for what its advertising should look like.
The competitor's founder looked like a fitness model and had his own influence. His ads centered on him. Our client's ads needed a reason for people to care about the piece on their own wall.
Copying the founder's ads couldn't reproduce his influence
The client copied the competitor's artwork style and tried to replicate the ads closely. I can understand the starting point. An ad already working in the same category gives a team something concrete to study.
The missing part was the relationship between that founder and his audience. His influence helped explain why people responded to him. A similar-looking ad from another brand arrived without that relationship.
We had our own account's history to examine. People had already seen the client's ads, responded to different messages and bought products. I used that history to compare how different groups responded.
That changed the question we were working on. We could look for which groups responded to our products, which creative they responded to, and where spending wasn't producing useful results.
The competitor's ads worked because a good-looking founder was in them talking about himself. Running the same angle for a brand that sells steel with a family name on it borrows the format and none of the reason it worked. Buyers wanted the piece itself, on their own wall.
Ad records showed which buyer groups performed well
I exported the ad data to CSV files and organized it in Google Sheets. Macros, calculations and a lot of pivot tables let me compare performance across creatives, copy and headlines, then examine groups by age, geography, device and placement. A cohort here means a group of records sharing a characteristic we can compare.
I also compared performance over time, using the history available for each campaign. A result from today could look different against the previous few days or a longer period. Some campaigns had enough history for that longer view. Others didn't.
Women aged 45 and over were a valuable group. iPhone users also performed well. Those findings gave us directions to test and places to reduce wasted spend. They didn't explain, by themselves, what a person wanted a family-name sign for.
We developed the creative around family and home. Normal families and ordinary settings worked well. One of the best-selling ads used a customer's photograph of a metal-art piece on his barn. The product was already in a place where someone used it.
"Family is Forever" and "Make Your House a Home" were among the angles we used. Gifts were part of the business, but most customers were buying for their own homes. Treating every buyer as a gift shopper would have missed that distinction.
I then tailored ads to the valuable groups and separated campaigns by age and gender, including a younger group. The analysis could now inform both what we said and how we divided the budget.
One blended campaign gets split into a campaign per age and gender band, winners get fed, losers get killed, and one band gets a deliberate push. Step through it: each bar's width is that band's share of the budget.
The competitor's ads sold on his own following, which a copy couldn't borrow. A customer's own photo of a sign on his barn became an ad, and the new ads kept selling, past $1M in about seven months.
Google Sheets was enough to make the comparisons
The tools were Facebook Ads, CSV exports, Google Sheets and manual macros. I didn't know Python at the time, so I did the analysis in spreadsheets. The work took many hours. I don't remember the exact calculations well enough to offer a reconstruction of the original workbook.
We could compare groups, avoid parts of the audience that were performing poorly, and brief ads around what was working.
The iPhone finding belongs to that account and that period. I used iPhone targeting as a starting point for some campaigns, then broadened the audience. My later account of the work also warns that changes to the data iPhones provided to Facebook made that old approach unsuitable as a present-day recipe. The case supports investigating a device difference. Current targeting decisions need current evidence.
Monthly revenue reached $150K in 90 days
The store started at around $10K a month and was barely breaking even. We reached $150K a month in 90 days. The business then passed $1M in revenue within seven months. That later figure describes revenue reached over the period, not a $1M monthly run rate.
These are results from my account of the engagement. I no longer have the original case write-up from the expired domain, and the figures here aren't an independently audited financial record. They also don't isolate the contribution of the copy, targeting or campaign structure. We changed those parts together.
The client later built an in-house team and continued to become an eight-figure company. That growth included its own operational expansion. My work helped establish the advertising approach it continued using. The client's later revenue belongs to a longer business history than this engagement alone.
The founder had a basis for creative and budget decisions
Separating campaigns by age and gender gave us control over the money allocated to each group. We could support promising campaigns, stop poor performers and decide when a larger test justified more budget. A larger campaign still involved risk, and I discussed those spending decisions with the client.
The founder also had a clearer basis for creative briefs and seasonal pushes. Family and home gave the team something specific to work from. Gifting remained relevant to seasonal campaigns without becoming the explanation for every purchase.
The business's later growth required more than keeping ads running. The client brought work in-house, added products and expanded its operations, including a move from plasma cutters to laser cutters. It continued using the campaign and product approach we'd developed while changing the business around it.
The offer was rebuilt around what the piece is for once it's on the wall. That's why the winning angles were about family and home rather than the product, and why an ad built from a customer's own photograph outsold the studio work.
A buyer group still needs a reason to buy
The age and device findings helped us decide where to test. Family and home helped us decide what to say. Those are related decisions, but an age bracket can't supply the message on its own.
The barn photograph made the intended use visible. A customer had put a piece of metal art on his barn. The ad showed the product in that setting.
Keep the observation and its interpretation separate. We observed differences between groups in the account. We developed messages around family and home, then used campaign results to decide what deserved more spend. A promising explanation still had to work in an ad before we increased the investment.
Start with the buyers your account can show you
This case is useful for a DTC founder whose store already has traffic, sales and enough history to compare groups. An export can give a creative brief a firmer basis when an account's overall result hides differences between audiences or messages.
The first useful question is what a comparison would change. A group performing poorly might justify reducing its budget. A family-centered ad performing well might justify another creative test. A weak or short history might justify collecting more evidence before making either decision.
A new store without that history needs a different starting point. This case also can't promise another store the same revenue or make an old device tactic current.
The family-name pieces gave us a specific place to start. We could examine who bought them, show the pieces in actual use, and fund the ads that worked. That's the work to reproduce in another account, with that account's customers and evidence.