『eCommerce Podcast』のカバーアート

eCommerce Podcast

eCommerce Podcast

著者： Matt Edmundson
無料で聴く

概要
 If you’re looking for great tips and insights into how to run your online store, look no further than the Ecommerce Podcast: a show dedicated to helping you deliver eCommerce WOW. New episodes are released every Thursday, and each episode features interviews with some of the biggest names in the eCommerce world. Whether you’re just starting out in eCommerce or you’re a seasoned veteran, you’re sure to learn something new from each episode. So what are you waiting for? Subscribe to the Ecommerce Podcast today!Aurion Digital マネジメント マネジメント・リーダーシップ マーケティング マーケティング・セールス リーダーシップ 経済学
エピソード
  • The AI That Calls Abandoned Carts for 65 Cents

    The AI That Calls Abandoned Carts for 65 Cents
    2026/11/12

    Andreas Filippov's AI calls customers the moment they abandon a cart, and it's recovering 15 to 20% of them for about 65 cents a call.

    Andreas Filippov joins Matt Edmundson on The eCommerce Podcast to talk about Loop Voice, the AI voice platform he's spent the last six months building. Loop phones customers who abandon their cart, works through whatever's holding them back in a live conversation, then texts a direct checkout link to finish the sale. Across more than 100,000 calls, Andreas and Matt get into the economics of the channel, whether customers mind talking to AI, and where voice fits (and doesn't) in a brand's wider marketing mix.

    Key points:

    06:54 Should Customers Be Told They're Talking to AI

    12:44 What 100,000 Calls Taught Loop Voice About Customers

    13:48 How the Abandoned Cart Call Actually Works

    31:08 Where AI Voice Calling Should Not Be Used

    Should Customers Be Told They're Talking to AI

    [06:54] Matt pushes on the one thing that bothers him about AI voice calling, that he feels duped if he's not told he's talking to a machine. Andreas's answer is that most brands don't lead with it either way.

    "Most of our customers don't really brand it as AI."— Andreas Filippov

    The agent introduces itself with a name, like "Lisa from Brand XYZ", and gets on with solving the customer's problem. Andreas says complaint volume about the calls has stayed low across more than 100,000 calls.

    What 100,000 Calls Taught Loop Voice About Customers

    [12:44] Beyond recovering the sale, Andreas has found a research use nobody built the product for directly. Founders who've shipped millions of units and never once heard a customer's actual voice can now listen back to real conversations.

    "We have a full conversation with the customer on what their pain point is, how they feel, at what point they thought, okay, I need to solve it."— Andreas Filippov

    Those recordings surface specific product, pricing and messaging objections that a tick-box post-purchase survey rarely catches.

    How the Abandoned Cart Call Actually Works

    [13:48] The mechanics are simple. Loop waits around ten minutes after a cart is abandoned, then calls.

    "We've seen you have an open cart, you haven't finished the purchase. Is there anything that came in between you and finalising the purchase?"— Andreas Filippov

    Once the objection is handled (sizing, shipping, a checkout hiccup), the agent texts a direct checkout link so the customer can finish in one tap. At scale, Andreas says 500 abandoned carts a month at a 15 to 20% recovery rate and a $50 average order value works out to roughly $2,500 a month in recovered revenue, for about 65 cents per answered call.

    Where AI Voice Calling Should Not Be Used

    [31:08] Matt asks the obvious follow-up question, could this be pointed at cold outreach, calling people who've never bought anything at all.

    "That would be a big issue in the ecommerce space doing that."— Andreas Filippov

    Andreas draws a hard line between calling someone who's already shown buying intent (an abandoned cart) and cold-calling a list of strangers, which he thinks damages trust in the channel for everyone.

    Today's Guest

    Andreas Filippov is a fractional CMO with seven years in ecommerce, starting as a media buyer in 2019 before moving into head of growth and fractional CMO roles for eight-figure Shopify brands including Moon Magic and Mother's Earth. For the last six months he's been building Loop Voice, an AI voice platform for Shopify brands that recovers abandoned carts, runs post-purchase upsells and handles inbound customer support by phone.

    • Website: loopvoice.ai
    • LinkedIn: Andreas Filippov

    Episode link: https://www.ecommerce-podcast.com/ai-abandoned-cart-recovery-andreas-filippov
    続きを読む 一部表示
    48 分
  • How to Get Your Products Recommended by ChatGPT

    How to Get Your Products Recommended by ChatGPT
    2026/11/05
    Jim Boudreau has studied SEO since 2001, and even he had to ask ChatGPT what was actually in the box.Jim is the founder of CataSEO, an AI tool built in partnership with Moz that optimises ecommerce catalogues for Google and for AI search engines like ChatGPT, Perplexity and AI Overviews. In this conversation with Matt Edmundson, Jim explains why catalogue-scale SEO breaks the old playbook. Once a store has hundreds or thousands of product pages, hand-crafted content stops being realistic, and copying the manufacturer's description starts costing visibility everywhere that matters. Jim walks through what actually moves the needle for AI search. That means tight metadata, structured data, alt text that describes the image rather than the product, and FAQ sections built to answer the questions customers (and AI) are actually asking. He also tells two very ordinary shopping stories, buying a Microsoft Surface and a pair of monitors, that show exactly what goes wrong when that groundwork is missing. Along the way, Jim and Matt get into using AI as a co-writer without losing the original thought that only comes from years spent working inside a category.[09:29] Why SEO Breaks Down Once You've Got Hundreds of Products[39:16] The Keyboard That Wasn't in the Box[15:19] Metadata, Alt Text and the Category Pages Everyone Skips[25:11] Writing With AI Without Losing the Original ThoughtWhy SEO Breaks Down Once You've Got Hundreds of ProductsJim has been in ecommerce since 1998 and studying SEO specifically since 2001, long enough to have heard every guaranteed-page-one SEO pitch going, and to know none of them worked that way. The real problem, he tells Matt, is scale. Writing a strong product description for a handful of SKUs is realistic. Writing one for a few hundred, let alone a few thousand, is not, and that is exactly the level of detail AI search rewards."When you have hundreds or thousands of products, it's impractical to do SEO on a product level." - Jim Boudreau, [09:29]That gap is why Jim built CataSEO. Rather than trying to win from the top down with pillar pages and landing pages, it works from the bottom up, product by product, filling in the metadata most stores leave half populated.The Keyboard That Wasn't in the BoxBuying a couple of monitors from the biggest electronics retailer in the US, Jim could not find which cables were included in the box, so he bought extra ones just to be safe. Researching a Microsoft Surface for a friend, he ran into the same problem on a bigger scale. Every product photo on the manufacturer's site and on the retailer's site showed the Surface with its keyboard attached. Nowhere did either site say the keyboard was sold separately."You can't find anywhere except ChatGPT that this thing doesn't include the keyboard." - Jim Boudreau, [40:32]If the biggest brands in the world cannot get something this basic right on their own websites, most stores still have work to do, and the ones that fix it are the ones AI ends up recommending.Metadata, Alt Text and the Category Pages Everyone SkipsAsked what store owners can do without buying any tool, Jim keeps it practical. Do not copy the manufacturer's description. Even a quick rewrite with ChatGPT is better than lifting the original word for word. Keep URL slugs and page titles tight. And fix the small things almost nobody gets right, like alt text."What an alt tag is supposed to do is actually describe the picture, not the product." - Jim Boudreau, [16:18]Category pages get the same treatment. Jim's advice is to keep products above the fold, since that is what customers came for, while still giving Google, Bing and the AI engines a real description to read rather than an empty template field. Blog posts matter too, as long as they answer a genuine customer question instead of reading like a sales pitch.Writing With AI Without Losing the Original ThoughtJim is drafting two books at the same time, both with AI doing a lot of the heavy lifting. His view is that AI can handle the research and the structure, but not the insight that comes from years spent watching a category change."It can go scrape the internet and come up with the 95% of knowledge. But it's not going to come up with that innovation, that perspective, that foresight that you do as an expert." - Jim Boudreau, [26:43]Used that way, Jim says, AI becomes a co-writer rather than a content generator, drafting from a detailed brief, his sources, his points and his own style."It's not doing it for you. It's doing it with you." - Jim Boudreau, [30:45]Today's GuestToday's guest: Jim BoudreauCompany: Studio 1119, Inc. (maker of CataSEO)Website: cataseo.aiLinkedIn: Connect with Jim on LinkedInEmail: info@cataseo.netAbout The eCommerce PodcastThe eCommerce Podcast helps you deliver eCommerce WOW with real talk about building online stores. Every Thursday, Matt Edmundson talks to experts and founders who have been in the trenches, built the stores, and learned the hard way, so you...
    続きを読む 一部表示
    49 分
  • Are You Running Ads to Products People Can't Buy?

    Are You Running Ads to Products People Can't Buy?
    2026/10/14
    Ecommerce Merchandising, and How to Stop Advertising Sold-Out StockMost stores are quietly leaking sales, and ad spend, to products customers literally cannot buy. In this episode Matt Edmundson talks merchandising with Leon Bart-Williams, co-founder and CEO of Akikumo, a merchandising-intelligence app for Shopify brands. Before Akikumo, Leon built the first shoppable-TV commerce experience at NBCUniversal and helped hundreds of brands scale on Amazon, so this is a proper operator conversation about the part of ecommerce almost nobody owns.SummaryMerchandising is the most overlooked driver of conversion, and it is far more than shop-window mannequins. Leon defines it as putting the right product in the right place at the right time in front of the right audience, with the right inventory to satisfy demand. It is the process of turning inventory into sales.The conversation works through where stores go wrong. Out-of-stock products sitting in the prime slots of a collection. Ads pointing at sold-out or partially sold-out products. A failure to understand which variant is actually driving the sales of a given product. From there Leon and Matt move into curation over volume, why more choice tends to sell less, A/B testing collections rather than just the checkout, and inventory planning that starts with sales velocity at the variant level. Practical throughout, with real brands named and a ten-minute daily habit any store can start tonight.Merchandising Is the Job Nobody in Your Store Is DoingMerchandising usually gets wrapped in grandiose language. Leon's definition is plain."Merchandising is really the process of taking your inventory and converting it to sales." — Leon Bart-WilliamsThe point most brands miss is that merchandising applies to the ecommerce storefront, not just the physical store. The first positions in a collection do the same job as the front window of a shop."Think of the first 10 slots of your website as the moment people walk through the door of a physical store and see a mannequin with nothing on." — Leon Bart-WilliamsOut-of-stock products in the first ten positions of bestsellers or new arrivals are the fastest, most common leak.The problem starts biting at around 400 to 500 SKUs, not tens of thousands, especially in apparel where size, colour and style variants multiply.It is almost always human error and broken team communication, not strategy. As Leon's co-founder Jared (previously a principal engineer at Figs and Chewy) puts it, most of ecommerce is just fixing human errors and self-inflicted wounds.Are You Running Ads to Products People Can't Buy?A costly and common mistake is ad spend hitting stock that is already gone. Leon describes checking a friend's brand, Johnny O, in Meta's Ad Library and spotting a top-selling product with three quarters of its variants sold out, sitting in the top ad slot. One email to their ads team fixed it.The subtler version is the partially sold-out product, and this is where the 80/20 rule matters at the variant level."The dress in blue, size 2, is driving 80% of the sales of this specific product. You need to know that, and you need to at least be fully stocked in that product." — Leon Bart-WilliamsRoughly 20% of a product's variants drive roughly 80% of its sales.If that core variant is out, the listing can look in stock while the ad is effectively pointing at a sold-out product.When a product is gone but the creative still works, point the ad to the collection, or to the category, rather than a dead page. When in doubt, point to the collection.Why More Products Means Fewer SalesLeon and Matt agree that curation beats volume, and the science backs it up. Matt recounts the jam experiment, where a display of six jams sold well and a display of thirty-six exotic jams saw sales fall. More choice created paralysis, not more buying."More does not mean more sales." — Leon Bart-Williams"People want fewer options. People want to be told what to do." — Leon Bart-WilliamsKeep collections tight. Around 50 SKUs or fewer per collection is manageable and tends to convert better.Curation is part art, part science. The science is how many products. The art is telling a story around a hero product, the dress plus the handbag plus the shoes, and writing content around it.Brands Leon rates for this include Cara Cara NYC, and his customer Brooke at Simply Moab and Desert Soul in Moab, Utah, who matches her ecommerce experience to her physical stores.Endless scroll with no curation is the opposite of this, and it kills conversion.Test Everything, and Plan Inventory at the Variant LevelTesting should not stop at the checkout. Leon argues for A/B testing collections, layouts, price points and new products, each with a hypothesis and a finite goal."You can't optimise your way to success all the time." — Leon Bart-WilliamsThe counter-example is Vince at the Texas golf brand Bogey Brothers, who built sale pages by hand from memory, guessed the pricing, ...
    続きを読む 一部表示
    49 分
adbl_web_anon_alc_button_suppression_t1
まだレビューはありません