I was the second bid ($2) on Outbid and I still get traffic from it… 😅 Thanks @jonathan_wilke 🤝

You can catch some broken addresses with Shopify Flow for free. Here's how, and why it's not enough. Trigger: order created. Condition: shipping address line 1 contains no digit. Action: add a tag like "check-address" and hold fulfillment. Then someone on your team emails the customer, waits for a reply and updates the order by hand. It works, sort of. But: It only finds missing numbers. A wrong house number, a typo in the street or a street that doesn't exist all pass. It flags too much. "Am Markt" or a PO box without a number gets tagged too. It happens after the purchase, so the fix still costs a support email and a delay. Many customers never answer. It doesn't work the same in every country, because some address formats put the number first or don't use one at all. So Flow tells you there's a problem. Someone still has to fix it. That gap is why we built AddressPill.

AddressPill runs on 7 Shopify Plus stores and around 25 regular Shopify stores. Together that's almost 500k orders a month going through our address check. We have 1 review. Why? The Plus stores came through our agency. They're clients, and I get why they don't write App Store reviews. The other stores set everything up on their own. Not a single support request. Which is great, except I never had a moment to ask for a review. The one review we have came from a store that wrote it without anyone asking. Still grateful for that. So the next goal is simple: more reviews. If you use AddressPill and it saves you time, a short review would mean a lot 🙃

Most Shopify address checks only look at the format. Is there a house number, is the postcode the right length. AddressPill also checks if the street actually exists. If it does, the order goes through. If it doesn't, AI looks for the likely typo and suggests the fix, fast enough that the customer barely notices. "Hauptsrtasse" becomes "Hauptstrasse" before the parcel ever leaves.

One of our Shopify Plus client is the reason AddressPill exists. About 60k orders a month. Around 7% came in with an address that couldn't be delivered as entered: missing house number, wrong number, typos. That's roughly 4,200 orders a month where support had to email the customer, wait for a reply, update the order, and sometimes ship twice. With everything included, each case cost them somewhere between $10 and $15. Do the math and it's a five-figure problem every month, just for addresses. So we built a check that catches this before the parcel leaves. Customers fix it themselves at checkout or right after the purchase. Today they pay less than a tenth of what those cases used to cost, and the errors are down to 0.13%. The boring problems are often the expensive ones.

Since May we have a new Shopify app live: AddressPill. It started because a lot of our clients kept getting orders with addresses that couldn't be delivered. Missing house numbers, wrong house numbers, typos etc. Every one of those means paying for shipping twice, CS and a customer asking where the parcel is. So we built an app that catches this before the parcel leaves. I'll share more about it here over the next weeks, including what doesn't work. If you run a Shopify store: what goes wrong with your deliveries most often?

Your SaaS should be visible in AI Search

We're expanding Genseo beyond AI visibility automation. Started as a tool to get your brand visible in AI search results. Automated everything. Set it and forget it. Now we're moving into the agency market. And agencies need more. So we're building: Prompt tracking. See exactly where you rank in AI search. Competitor tracking. Monitor where competitors show up. Model-specific insights. Track your position across different AI models. ChatGPT. Claude. Perplexity. All of them. Performance overview. One dashboard. All your AI visibility metrics. Why this matters: AI search is the new SEO. But nobody knows where they actually rank. You optimize content. You create answers. But you're flying blind. Not anymore. You'll see: - Which prompts trigger your brand - Which models rank you higher - How your visibility changes over time - Where competitors are beating you Agencies managing multiple clients need this visibility. You can't optimize what you can't measure. Genseo now gives you both. Automation + tracking. AI visibility is growing fast. Time to track it like you track SEO.

ChatGPT almost stopped citing Reddit overnight. That’s a pretty important lesson for anyone trying to “optimize for AI Search”. For a while, Reddit was treated like one of the most important sources for getting cited by ChatGPT. Then on August 14, Reddit’s share of ChatGPT Search citations dropped below 1%. The 7-day average fell by more than 50%. That’s the part of AI Search I think many companies underestimate: The rules are not stable yet. A tactic that looks incredibly important today can lose relevance a few weeks later. Which means building an AI Search strategy around one source, one platform or one optimization trick is probably a bad idea. The real goal should be much broader: Make your brand easy to discover, understand and trust across the web. * Your own website. * Industry publications. * Reviews. * Communities. * Comparison pages. * Directories. * Other authoritative sources. Because we’re not optimizing for Reddit. We’re optimizing for the systems that decide which brands deserve to be part of the answer. And those systems are going to keep changing.

If ChatGPT recommends 5 competitors and not you, your problem probably isn’t ChatGPT. It’s tempting to think there must be some AI Search trick you’re missing. A prompt hack. A certain page structure. A specific platform you need to appear on. Sometimes that may help. But usually the bigger question is: Why is the internet giving those competitors stronger signals than you? Maybe they’re mentioned more often by trusted sources. Maybe their positioning is clearer. Maybe more people review them. Maybe their content answers the exact buying questions customers ask. Maybe the category association around their brand is simply stronger. AI models don’t operate in a vacuum. They learn from the information available around your company. So if your competitors keep getting recommended, I wouldn’t start by asking: “How do we optimize for ChatGPT?” I’d start with: “Why is our competitor the more obvious answer?”

I asked ChatGPT to recommend companies in several different categories. The companies it DIDN’T mention were more interesting than the ones it did. Some of them had: Strong brands. Good websites. Thousands of Google rankings. Years of content. And still... When someone asked ChatGPT for a recommendation, they basically didn’t exist. That’s a pretty strange situation. A company can spend years becoming visible on Google and still be invisible in a completely different search experience. I think this is one of the biggest changes marketers will have to understand over the next few years. The question is no longer just: “Where do we rank on Google?” It’s becoming: “Where does our brand appear when someone is looking for a solution?” Google. ChatGPT. Gemini. Perplexity. AI Overviews. And probably platforms we aren't thinking about yet. I’m going to spend a lot more time testing this. Not because I think Google is disappearing. But because search is becoming much bigger than Google.

POV: Some people think I bid on Outbid.lol because of the platform. But why don’t people understand that all bidders only do this because of the audience???

What's the best AI for generating realistic B-roll footage? I need actual, usable B-roll. Not weird AI artifacts. Not uncanny valley stuff. Something that looks clean. Professional. Like real footage. I've tried a few tools, but the results are hit or miss. Either the motion is off, or the details look fake, or it just doesn't hold up. What are you using? What actually works? Drop your recommendations.

Haven't posted here in a while. Been heads down rebuilding Genseo's onboarding from scratch. We just shipped v3. And honestly, this is the part most founders ignore. Everyone thinks onboarding is just forms and questions. Something you build once, ship it, and move on to the "important stuff". I used to think that too. But here's what I learned over the last months: Your onboarding determines everything. It's not a nice-to-have feature. It's the foundation of your entire product. If your AI doesn't understand the user's brand, audience, and positioning from day one, it can't deliver results. Garbage in, garbage out. What seems "unimportant" on the surface is actually the difference between content that ranks and content that disappears. Here's how our onboarding evolved: → V1: Manual forms. 20+ fields. Users had to describe their brand voice, list competitors, define their audience. Half got it wrong. The other half left fields empty. → V2: Better UX, cleaner design. Still required too much manual work. Better than v1, but not good enough. → V3: You paste your website URL. We extract everything. Brand voice, target audience, positioning, competitors. You just review and confirm. A few fields instead of 20. The result? Better data accuracy. Better AI output. Better blog performance. Higher rankings. The boring details are the ones that decide if your product works or fails. If you're building AI products, don't treat onboarding as an afterthought. It's where quality starts. PS: If you're running a brand or building a SaaS, curious what you find most frustrating about onboarding experiences. Let me know in the comments.

I used to check Stripe every hour. Sometimes multiple times per hour. Refresh. Check MRR. Refresh again. Check churn. Refresh. Exhausting. Now I use ChartMogul. They send me a daily update. One email. All the metrics I need. MRR. Growth. Churn. LTV. Everything. I still care about the numbers. But I'm not obsessing anymore. The data comes to me. I don't hunt for it. If you're stuck in the refresh loop, get a tool that sends you updates. Saves time. Saves mental energy. You still stay informed.

Anyone know if Loops.so supports MCP yet? We're using Loops in 2 projects right now. Compared to other email platforms, setting up transactional emails is painful. Every single email needs manual configuration. Templates. Triggers. Data mapping. It takes forever. Other platforms we use? Way smoother. Faster setup. Less friction. @loops is solid for what it does (the best), but the workflow is exhausting. If they had MCP integration, this would be 10x easier. Has anyone heard anything about this? Or found a better way to handle it?