Tom Luce

Now that we have LLMs, bulk exports out of Screaming Frog are an absolutely delicious move. One of my favorites is "Screenshots". (Rendering ON, Store rendered HTML) Toss them into a local folder and point Claude Code to it. I like to collaboratively audit: - Design (the look) - UX (the experience across pages) - CRO I also love digging into the page templates - the sections that make up key page types - to ensure they're as effective as possible. For smaller sites you can look at everything. For larger sites, I'll focus more on key pages, like home and money pages. When you settle on some recommended updates, you can draft pages in the style of the client's site with Claude's front end design skill (find it in the marketplace). Pro tip: Tell it to design with the site's current, perceived design capabilities in mind. In other words, don't do anything that seems like they wouldn't reasonably be able to implement, based on what it sees. Get your bulk export on!

85

9 Comments

The Death of the "Granular" Performance Marketer 🪦 Stop micro-managing your campaigns. You're killing your ROI. 🛑 If your Google and Meta accounts still look like a patchwork quilt of 50+ hyper-segmented ad sets and "exact match" silos, you aren't being "thorough." You’re being an obstacle to your own growth. The era of the "PPC Plumber" is over. We have entered the era of The Aggregator. Both Google and Meta have made their stance clear: Aggregation is the new best practice. Why? Because AI needs data volume to breathe. When you slice your budget into 100 tiny segments, you starve the machine. Here is the new 2026 performance playbook for those who want to lead the industry: 1. From Granularity to Aggregation 📉➡️📈 Stop trying to out-think the algorithm on "who" sees the ad. Consolidate your structures. Use Broad Match (Google) and Advantage+ (Meta). By aggregating your conversion data, you give the AI the statistical significance it needs to find your next customer at the lowest possible CPA. 2. Creative IS the New Targeting 🎨 In a world of broad audiences, your creative does the heavy lifting of filtration. The old way: Targeting "Men, 25-34, interested in hiking." The new way: Using a video that speaks so specifically to a hiker’s pain points that the AI learns who to target based on who stops the scroll. Your job is no longer to find the audience. Your job is to attract them. 3. Feed the Beast (Signal over Setup) 🤖 The most successful marketers in 2026 aren't the ones with the best manual bid adjustments. They are the ones with the best first-party data signals. If you feed the AI high-quality conversion data, it will work for you. If you feed it junk, it will scale junk. The Bottom Line: Shift your time away from "pulling levers" and toward Creative Strategy and Data Integrity. That is where the alpha is found today. Are you still holding onto granular control, or have you fully embraced the "Power of the Machine"? Let's debate in the comments. 👇 #PerformanceMarketing #GoogleAds #MetaAds #AI #GrowthMarketing #DigitalAdvertising #CreativeStrategy

5

6 Comments

Most multi-location health and wellness brands think they have a “media performance” problem. What they actually have is a visibility problem. We recently audited a healthcare group group spending $2.8M across Meta and Google. Their location-level data looked fine on paper—but when we isolated spend and lead quality by market, over 42% of their budget was driving zero attributable appointments. Why? Their systems couldn’t connect local campaigns to real conversions. Each location ran their own ads, each with different tracking setups, landing pages, and creative. Once we ran our Meta + Google Audit and centralized attribution, the brand: 1. Identified $500K in wasted spend 2. Reallocated to top 20% performing markets If your locations aren’t growing evenly, it’s probably not the leads—it’s your visibility layer.

6

1 Comment

If I were to start a brand today, I wouldn't look at Google Analytics, platform ROAS, or really ANY granular "performance" metrics. Including any kind of "incrementality" metrics or other complex performance (MMM, MTA, inc testing, etc). I'd simply look at: • Daily trending P&L metrics (CM) • Daily trending new customers (NTF) • Daily trending efficiency metrics (CAC) • Monthly cohorts of customer value (LTV) • Eventually get to a P&L view that has cash flow Every day I'd be able to see the last 7, 30, 90, and 365 trends of the above four metrics. I'd actually do a rolling average of each of those time periods because I don't believe in day-by-day unless I'm super tight on cash. How do I know what's working and not working? • List of hypotheses and tests • List of actions taken + tests done • Change of BUSINESS kpis (above) after each • Investigative exploratory data analysis of unexpected changes For example, I'd have something like: August • Problem: CAC is too high • Hypothesis: spend is too high • Test: reduce overall ad spend by 20% • Results: monitor CAC trend over next 30 days In September, we saw that CAC was reduced by 10% vs the trending 30 day average. This got our CAC under control and was deemed "successful" as a test. To get more granular, I might have a hunch that a particular channel or tactic is less effective at adding incremental customers. Rather than a blanket 20%, maybe we cut 50% out of Google Ads and keep everything else flat, etc. You could have tests surrounding other areas of the business beyond just ad spend: • New product drops • Pricing strategy • Promos • PDPs • New ICP • New category • New overall channel Etc. And then you'd have to keep your tests fairly clean, testing only 1 directly integrated activity<>KPI relationship at a time. For example, a lot of things affect new customers. So you need one "chief" test at a time that takes a big swing. Keep in mind -- I'm talking small brands here. If you're working on things and overall your brand is growing on those primary KPIs every day or month (e.g. contribution margin, new customers, etc) then you're in a good state. If you're not, then you need to be more prescriptive with those tests. Once you're at like $50-$100m a year, then you can start getting cute with measurement. MAYBE even $25m a year. But if you're sub $20m, quit wasting time. (For the record, there are clients at Power and fusepoint that are multiple hundreds of millions of dollars just focused purely on CAC and iCAC, at least for their marketing teams). How would you run a startup? #attribution #cac #startup

124

8 Comments

Alright, back from Snap Inc.'s performance summit. Here’s my take: In a nutshell, Snap is putting in the WORK. They’re actually building the things everyone's been yelling about: --> Better measurement (CAPI, 3rd-party stuff)--all the cool data things I grasp broadly but otherwise attribute to the work of smarter people who know more than me --> Automation nation--budgeting at the campaign level has arrived! --> New ad formats for DR--sponsored ads, Snap's answer to whitelisting/partnerships that shows up in users' chat list (very cool, very effective) We know the playbook is different. It's simply not Meta/Instagram/TikTok. User behavior, audience demographics--it's almost as strategically dissimilar as YouTube is. Regardless of your degree of skepticism, one thing is for sure: they are not stopping until they convince you of the progress they've made. I personally will continue to advocate to test and retest, so long as the changes are substantive (and the incentive is there, of course 😉). To all the advertisers side-eyeing Snapchat: what are you thinking about?

49

What's the best path to bringing search and social budgets to CTV? That’s one of the questions we debated on the latest episode of Open Market. Joe and I were joined by our friend James Borow of Comcast's Universal Ads for a wide-ranging conversation that hit on: - The state of the CTV market (and why it’s still early innings) - Universal Ads’ big bet: building an API-driven platform for what James calls "economically rational" marketers - Lessons from building and selling two startups: what James would do differently, and how those experiences shape how he’s leading today - The power of compounding in careers: how relationships, reputation, and knowledge stack over time, especially in adtech James is one of the sharpest thinkers in our industry. You're going to love this one. Link below.

53

12 Comments

When Facebook, Meta, and Amazon promise to manage campaigns end-to-end from creative to targeting to measurement...no one seems to question that they’re doing so within media ecosystems they themselves own. Good news for the holding companies: if they keep trying to compete with the big three while playing at a disadvantage, they likely won’t be around in 10 years.

22