Which Shopify Apps Actually Drive Revenue?
One affiliate. Twelve months.
EMC Podcast Episodes
Are Blogs Still Relevant for Ecommerce Stores in 2026?
Most ecommerce blogs fail for one reason. somebody wrote about what they felt like writing about....
Should you hire a freelancer or an agency?
Both hosts are freelancers and say so at the top of the episode. Jimmy's case against agencies is turnover and seniority: you get a new account manager every few months, and unless you are spending six figures a month you get a junior learning on your budget. The honest counterweight is that a freelancer has no bench behind them, some have only been going a few months, and a case study result does not mean the person who produced it will ever touch your account.
What should you be doing for Black Friday right now?
Decide the offer first, then build everything around it. Jimmy's timeline is September to plan, October to set up and November to run, and his strongest point is email: one brand has 48 campaigns scheduled between 1 November and 7 January, against the two most stores send. The offer does not have to be a discount, and he points to Merry People, who never discount and give away a pair of socks instead. Charles's side is to get the Black Friday landing page live now, so it has time to rank before anyone searches for it.
Should an ecommerce store owner start a podcast?
Both hosts say yes, and the episode doubles as them rethinking their own show. One recording becomes clips, shorts and ad creative, and syndication to Apple, Spotify and YouTube produces genuine backlinks to your store as a side effect. Their own lesson, said out loud here, is to cover one specific topic per episode rather than a weekly roundup, because a single topic is what people search for and send to each other.
Which numbers do you need before you can tell whether your ads are working?
The break-even return, not the ROAS showing in the ad account. Jimmy's point is that a 10x return can still lose money once warehouse costs, staff and the rest of the overhead are counted, and plenty of owners cannot say what return they actually need to be profitable. He will not take a client on until that number exists, because without it nobody can judge whether the account is doing well.
Why does going after a broad niche kill a new store?
Because you have no authority to compete at the top of a category and nothing specific enough to win at the bottom. The hosts' answer is to pick a narrow sub-niche and a defined demographic, win there, then work up to the broad category once you have reviews and credibility. Jimmy also covers the Merchant Center price benchmark report, after a student matched the benchmark price on a $2,000 product and sold one within five minutes, having sold none for a fortnight.
Why is relying on one traffic channel risky?
Because one lever is not a moat. Jimmy checks attribution in client accounts and treats any store taking nearly all its traffic from a single platform as fragile, since there is no room to fix a problem when the only channel breaks. Charles frames the healthy version as three pillars, paid, organic search and organic social, with email on top for the list you already own.
What did one founder do differently to go from 30K to 150K a month?
Execution speed, not a better product or a bigger budget. Jimmy coaches the founder in question, who joined the academy in February on roughly 30K a month and reached 150K. The thing that separates him is that he takes a system, implements it immediately, and improves on it. Jimmy's blunt assessment is that this is a level ten operator running a level three product, which is why the numbers moved so fast.
Why has AI not given you any time back?
Because the time saved gets refilled rather than kept. Jimmy saw a LinkedIn poll asking who actually has more time since adopting AI, and almost everyone answered that they were as busy as ever. He puts himself in that group: he does two months of work in two weeks and takes none of it back. His mentor's diagnosis was that he never safeguarded the time, so the work simply expanded to fill it.
If you turned off your ads tomorrow, would you still have a business?
For a lot of stores the honest answer is no, and that is the risk worth naming. Jimmy brings back the example of a founder who launched by going heavily at SEO, building dedicated pages around phrases doing as few as 70 searches a month, and had organic traffic before paid ever carried the business. The hosts work through the channels available beyond ads, including blog content, YouTube, community and the podcast itself.
Why did Charles leave Semrush for SE Ranking?
Because everything had become an add-on. AI Search visibility tracking cost extra, API access to link up Claude cost extra, and the total stopped being worth it against what else is available. Charles moved to SE Ranking on the grounds that it offers more for less. The rest of the episode covers Claude Design, which he has used to win a client from a proposal and to build a design system per client.
What is it actually like running ChatGPT ads?
Jimmy's first finding is logistical. Ads sit in a pending approval queue for around two weeks, so start the process before you intend to spend. ChatGPT recommends a cost per click of three to five dollars, which is high for most ecommerce, and you have to add your own UTM parameters because the native tracking is hard to follow. Targeting is set by describing the conversations you want to appear in rather than by keywords alone, and nobody has enough data yet to say which approach wins.
What does Google's AI search update mean for your store's traffic?
AI Overviews and AI Mode now sit alongside each other in the main search experience, and Google reported 2.5 billion monthly users on Overviews with AI Mode at a billion and doubling every quarter since launch. Charles is seeing impressions rise year on year across client accounts while clicks trend down, because the answer is given in the interface. The job shifts from capturing clicks to capturing visibility.
Does SEO still matter for AI search?
Yes, and Google has now put it in writing. Its documentation says generative features are rooted in core search ranking and quality systems, so the same best practices apply, and that optimising for generative AI search is still just SEO. Charles's caveat is that what Google says and what actually works are not always the same thing, and platforms outside Google have their own quirks.
Why build your own Claude skill instead of copying someone else's prompt?
Charles built one that handles SEO titles, descriptions and keyword research with access to his rank tracking data, so it picks a target keyword by search intent rather than by volume. It got the first metadata pass done across a large catalogue site in weeks instead of months, with him reviewing each section of the navigation as it went. A skill encodes how you actually work, which a prompt you copied off LinkedIn cannot.
Which AI model should you use for which task?
Jimmy's answer is to stop picking one and track arena.ai/leaderboard instead, which ranks models by task and moves week to week. At the time of recording Claude led on text and reasoning while ChatGPT's new image model had taken top spot for generating and editing images. The second half covers real AI referral data, including one client taking around 30 orders from ChatGPT and Bing against 77 from Google organic over a fortnight.
What does a website redesign really cost you when it goes wrong?
One of Jimmy's clients spent $12,000 on a redesign and lost more revenue the following month than the redesign cost. Nobody told the SEO or Google Ads freelancers it was going live, so the GA4 tracking code was stripped out, days of Ads data disappeared, and homepage rankings tanked because the new design removed the text that was ranking. Tell whoever runs your marketing before the site changes, not after.
What can Claude Code actually build for an ecommerce store?
Jimmy built a tool that reads the misleading health claims report out of Merchant Center, finds every flagged product, and rewrites the descriptions to be compliant. It handled 600 products in 23 minutes against a job the client had been putting off for a year. The wider point is that large catalogue stores can now move at the speed of a small one, though both hosts warn you to work on a duplicate theme, because people have already taken live sites down.
Claude or ChatGPT? Which should ecommerce marketers use?
Both, for different jobs. Jimmy's view is that Claude is well ahead for marketing work and automation right now, and Charles cancelled his paid ChatGPT plan because he had stopped opening it. But ChatGPT is still the single biggest AI referral source in his clients' analytics, so it matters as somewhere customers search even when it is not the tool you work in.
Should you act on AI marketing advice without checking it?
No, because the output is usually close rather than correct. Charles uploads Search Console exports to Claude for year over year collection page analysis and still reviews every suggestion before actioning it, which is where an SEO still earns their place. Jimmy's related point is that lean stores now move faster than large ones, not because their tools are better but because they have no approval queue.
How can a solo ecommerce founder use AI instead of hiring?
Charles runs AI agents on a separate Mac Mini with its own Apple ID, so they never touch his personal accounts, and uses Claude for work that used to take hours: monthly expense reconciliation, competitor analysis across five sites, and backlink gap reports built from a Semrush export. Jimmy's starting point is simpler. Screenshot a high converting site in your niche, screenshot your own, and ask Claude to draft the rebuild.
What can you actually afford to pay per click?
Jimmy built a calculator that answers it from three numbers: average order value, conversion rate and profit margin. Charles covers the new Bing Webmaster Tools AI dashboard, which for the first time shows which of your pages Copilot is citing and which queries triggered those citations.
Which AI tools are actually worth testing in 2026?
Charles covers Claude's browser extension for writing product descriptions and running competitor UX analysis inside Shopify, and SEO Gets as a visual replacement for Search Console. Jimmy points at the Shopify marketing report most owners have never opened, which shows exactly which products are being bought through ChatGPT and other LLMs.
Why do generic keywords cost you money?
Generic terms catch people researching and specific terms catch people ready to buy, and most stores spend against the first while hoping for the second. Charles uses budget first aid kits against premium MOLLE hiking kits to show the difference, and answers the SEO is dead line with data showing Google still drives over 95% of organic traffic.
Should ecommerce brands move early on ChatGPT ads?
Jimmy's view is yes, because the first mover window closes as soon as larger advertisers adapt. The more useful number in this episode is the gap between stores: one client takes 5% of its traffic from ChatGPT while another, running generic descriptions, got six visits in 90 days.
What does Google's AI Business Agent mean for your product descriptions?
The agent answers customer questions and recommends products on your behalf, and it can only work from what you have written. Jimmy's client data shows AI Search going from 0.5% of revenue at the start of 2025 to between 5 and 10% for some stores, and the ones getting that traffic wrote their own descriptions rather than using the manufacturer's.
Is Black Friday really not right for your brand?
Almost never, and both hosts treat that line as an excuse rather than a strategy. They cover what premium brands that refuse to discount can run instead, and Charles shares the push notification setup that produced over $100,000 in revenue on roughly two hours of management a month.
When did you last update your brand search ad copy?
Most brands duplicate their organic listing in their brand search ads, which wastes the slot entirely. Jimmy covers what to run instead over Black Friday, and Charles shares meta title research showing 57% CTR improvements from simpler wording, plus why collection pages and product pages need opposite treatment.
GEO, AEO or LLMO? What should you actually call AI Search?
Charles makes the case for AI Search over the acronyms, because searching for GEO services returns environmental consultancies and confuses the people you are trying to reach. Jimmy covers the question he gets asked most, which is when to scale, and why it should be answered with a break-even calculation rather than a ROAS target.
What happened when Google removed the num=100 parameter?
Google removed the parameter that let tools pull 100 results in a single request, which broke rank tracking across the industry overnight. Charles explains why most tools can no longer see reliably past the top 10, and what that means for the reports you have been sending clients.
How do you launch a premium supplement brand against Athletic Greens?
Hursh Dodhia-Shah sold his marketing agency and built Optiventure after a 35 day Himalayan expedition showed him a gap in adventure nutrition. He reached a 25% email conversion rate from only 500 subscribers by emailing every single buyer for research, and partnered with a content creator rather than building an audience from scratch.
What separates ecommerce founders who scale from those who plateau?
Jimmy breaks down five traits he saw in a client who went from 30K a month to nine figures in five years. The short version is that they decided quickly, reinvested instead of extracting, and treated six months of flat revenue as a warning rather than a normal patch.
The show ran as a YouTube channel for 37 episodes before the podcast feed launched in September 2025. Those earlier episodes stay onYouTube.