Ben sat down with the Live Local Marketing Podcast to talk about what sits behind a Google Ads account that scales. The conversation covers how much to spend, which brands we turn away, how a ROAS target should be set, and the sports brand that went from a single brand campaign at around £10 a day to nearly £100,000 in a month.
The full episode is below. If you'd rather read, the main points are written up underneath it.
If you take one thing from it: your ROAS target should come from your margins. A brand that asks for a 3x ROAS can be profitable at 1.3x, and the gap between those two numbers is most of the room you have to scale.
How much should you spend on Google Ads?
The honest answer is that it depends on the niche, and the research comes before the budget. Ben recently put together a rough strategy for a prospect where the average cost per click was around $50. Many of the ecommerce brands we run sit at $1 to $2 a click, because there's less competition or the product price doesn't justify bidding higher.
Those two businesses need completely different budgets to learn anything. That's why Keyword Planner is the first step. Pull the CPCs and search volumes, then work out what the spend needs to be.
Sometimes the research says no. Brands have come to us and we've told them Google Ads won't work at their margins, their product price or in the markets they sell to, because there isn't enough search demand to make a difference.
Margin and mentality
In the first months of the agency, we took on any ecommerce brand that came to us. That's normal for a new agency trying to grow. It also taught us what to screen for.
Founder mentality is the first filter. The founder has to be building a business for the long term. The clients who were in it to make money fast were the hardest to work with, and the performance showed it.
Unit economics are the second. We go through margins on the first call or during discovery. Brands have come to us making about $10 profit per order, and that makes paid search very hard. It only works if they're willing to run at a loss while the account gathers data and the landing pages get optimised.
"If the margins are healthy and the founder is motivated to build something real, the relationship is better, the performance is better long-term, and they make a lot more profit in the process."
Ben, on choosing clientsBuild the ROAS target from your numbers
At the agencies Ben worked at before JB, ROAS targets were guesses. A founder would pick a number based on what a friend's brand hits or what another agency said, with no link to their own costs. A target set that way tells you nothing about profit.
We built a unit economics calculator that we send to brands during discovery. It asks for the product price, the cost of goods, shipping and what the 3PL or fulfilment partner charges, then works out two numbers: the break-even ROAS, and the target ROAS for the profit margin the brand wants.
That changes what's possible. A brand that insists on a 3x ROAS caps its own spend. If 1.3x is actually profitable, everything between 1.3x and 3x is room to spend more, scale harder and drive more total profit.
Asked which single piece of data he values most, Ben came back to the same place. If we know what it costs to get an order from the checkout to the customer's front door, we know what we can afford to spend to win it. Break even at 140% and return 200%, and the account has headroom to scale, test new campaigns or push harder on what's working.
We hold ourselves to the same standard. The JB portal tracks P&L per client and per employee, and having that visible has changed how we make our own business decisions.
An AI system that keeps everyone in the loop
We've built our own MCP server and data warehouse, and the aim is to connect it to every tool the agency uses. For research, it has access to Keyword Planner and DataForSEO, so search volumes and CPC data for an audit come back in seconds.
The part the host was most interested in is the brand context. Every client has an external Slack channel and an internal one where the team discusses smaller changes, issues and wins. Every night at around 10 or 11pm, the system reads those channels, condenses what's useful and saves it to the warehouse.
The result is a live, evolving record of every account. A new team member, or our operations manager who isn't in Google Ads day to day, can ask ChatGPT, Gemini or any LLM connected to the MCP a question about a client and get an answer based on what's actually happened. Nobody gets left out of the loop.
Remarketing and the abandoned basket
Google Ads is a cold channel in most people's heads, but there's a remarketing layer in every account we run. Performance Max reaches warm audiences heavily on its own, and we add remarketing display and YouTube campaigns on top.
For D2C brands, where order values aren't huge and the sales cycle is short, this matters less than it does in SaaS. Ben has run Google Ads for SaaS companies with three-month-plus sales cycles, where a planned sequence of ads leading up to the sale is the whole game.
Where remarketing earns its place in ecommerce is in two moments:
- The second product. If someone has bought product A and product B is the natural accessory, they need to know it exists. You already have their email, and you can reach them on Google too.
- The abandoned checkout. Shopify's integrations with Klaviyo and Google Ads let you build audiences from people who started a checkout and didn't finish. They had the intent once. They got distracted by work, the washing machine or dinner, and a reminder is enough to bring them back.
Search terms as product research
Search term reports show what people want, including things the brand doesn't sell yet. When we see accessory searches, or people asking what goes with a product the brand already stocks, we take it to the client with volumes and an estimate of the conversions and sales it would bring in.
Clients ask us for this directly, too: keyword research to see which new products would work on Google before they commit to making them. Most welcome it, and several have gone on to test the product.
Video: snappy product shots and a Shopping carousel
Ben was candid that we don't run as much video as we should. Google Ads agencies, us included, focus on search and product feeds. That's changing. We're launching more Demand Gen campaigns, and clients are asking for YouTube Shorts ads.
Most of our clients are Meta-first brands, so the assets already exist. They're working on Meta, and the job is to make them work on Google. As top-of-funnel, they're powerful, and the search and Shopping campaigns pick up on the back of them.
For ads, short-form wins. What works best for our jewellery, fashion and sports brands is fast product footage. Run it through Google and the Shopping feed appears as a product carousel underneath the Short, so someone sees the product on a person and taps straight through to it.
Long-form has a different job. YouTube content has always helped search rankings, and longer, conversational queries in AI Overviews can pull from a video transcript as readily as from an article. For SEO and AI visibility, long-form is the right call.
Let the search terms write the content
We build thematic campaigns. A sports brand that sells shorts gets a running shorts campaign, and the search terms coming off it show what people want to know: how to wash them, how to use them, which ones to pick for a specific run.
Those questions become content. Part of it runs as ads, and the rest belongs on the product page, which has a direct cost benefit. Google scores how relevant your landing page is to the keyword, and that score feeds into what you pay per click. If people keep asking a question and your page doesn't answer it, you're paying more for every click than you need to.
When the data says the brand needs work
Some brands come to us having rushed the fundamentals. The SEO was skipped, the branding feels off, or the site is full of AI-generated images. If fixing it improves conversion rate on Google, we'll help. We've built landing pages for clients, and we audit websites and landing pages as part of the work. That can be a formal audit or a Loom video every week or two whenever we spot something, and over time it adds up.
Pushback still happens, mostly with legacy clients from before we defined who we work with. The hardest cases were brands where Google kept disapproving ads, or wouldn't run Shopping at all, because of issues with the website, branding, positioning or business information. They performed well on Meta and couldn't see why Google was stricter. Google's misrepresentation and Shopping policies are tighter, and no amount of account work gets around them.
Having a clearly defined ideal client has made those conversations rare.
From £10 a day to nearly £100K a month
The episode is named after one client. It's a sports brand selling mainly running sunglasses. When they came to us, the whole account was one brand campaign spending around £10 a day. In the month before the recording, the brand made nearly £100,000. They've since added clothing, but sunglasses are still the core of it.
Ben's view is that the founder is a big part of why it worked. They go to their factories, design the products themselves and build them around customer feedback. A founder putting in that much effort is the kind we want to work with, and it ties straight back to the first filter.
"A brand that puts in that much effort, and a founder that has that drive, is always going to be amazing to work with."
Ben, on the sports brandThe full numbers, account structure and timeline are in the sports brand case study.
Watch the full episode
The whole conversation runs 22 minutes. You can watch it on YouTube, and you'll find more episodes on the Live Local Business channel. Thanks to the Live Local Marketing team for having Ben on.
If you already run Google Ads, the discovery call can be an audit of your account. If you don't, it's an hour-long strategy session where Ben walks you through the opportunity. You can also message him on Instagram at @ben.ppc.
Quotes are lightly edited for clarity.