Mike Stoop (@ScaleWithMike) header
Mike Stoop (@ScaleWithMike) profile picture

CGO & Founder. I run CRO for DTC brands and build AI agents & workflows. Sharing what wins, what loses, and why.

Netherlands
Joined January 2025
Photo by Mike Stoop (@ScaleWithMike)Photo by Mike Stoop (@ScaleWithMike)Photo by Mike Stoop (@ScaleWithMike)
one visitor from chatgpt’s product cards was worth 5 visitors from paid social on one of our stores. $13.01 vs $2.61 revenue per visit, last 90 days those are the little product boxes chatgpt shows in its answers. 6.5% of the people who clicked one bought. email didn’t get close on that store 8 of the 12 stores i checked get none of it. since tuesday you can try to get in yourself: ask chatgpt what your buyer would type right before buying something like yours, and see if you show up or a competitor does claim your store on openai’s new business page, link in the reply. if you’re on shopify you can connect your products right from there then fix the product info on your best sellers. start with price and stock, chatgpt answers from that it’s one store and 51 orders so don’t treat it as gospel. anyone else seeing this in their ga4?
1
7
Fable 5.2 created a MIRROR EARTH as a game. Go anywhere. Destroy anything. Weather and more match the real Earth, in real-time. Go explore: spawn.co/@mattshumer/earth?r…
91
120
50
2,370
404,215
the weather part is what gets me. that's a lot of systems talking to each other from one prompt
21
Bazzly is having growth that I have never seen before fully organic, driven by AI mentions, and not something we truly control truly insane imo this is driven by the quality of the product more than the distribution effort itself (actually partially true as we use Bazzly to push Bazzly so it's also part of the distribution effort - i need to post more about how we use it to push itself)
this guy had 7 years of nothing he had a $7k/month job, a product, and $300 of MRR .. and he just sent me a DM replying to one of my post Filip built something very good but it was missing the last 10% that makes a product go from "good" to "great" so we made a deal: he builds, we advise (with Tom) and I distribute the building took way more time than expected we wanted to have the best infrastructure to be the best Reddit-based acquisition platform - and this is tough and I think we made that I am super proud of what we created with Bazzly Filip is gonna be rich and this is very well deserved 👏
85
8
6
162
17,487
fully organic is the best kind. what do you think is driving it most, word of mouth or search?
1
for the first time ever, I'm sharing everything I do across all my SaaS to rank on Google and get mentioned by AI every SEO and AEO lesson I've learned is here: 1. we show up in Reddit threads. AI pulls a lot of its answers from Reddit, so Bazzly finds the threads that are fresh, and even the ones already ranking on Google, and gets our products into the top reply 2. we get our products listed #1 in listicles that already rank on Google, with Outrank. when someone asks ChatGPT for the best tool, those are the pages it reads 3. we turn Reddit questions into blog articles. Bazzly shows us what people actually ask, and we feed those questions to Outrank 4. I regularly sit down and find new niches where our products could be useful. I research each one and feed it to Outrank, and Outrank finds new ones on its own too, so our keyword plans keep expanding into new markets 5. we publish an SEO article on our own blogs every day, also with Outrank 6. Outrank helps us refresh old articles regularly. AI and Google both prefer recent pages, so every page has to stay current 7. we build new backlinks every week with Outrank, and every placement gets verified. we also pitch guest posts using Outrank's free guest post database of 278 checked sites, all DR 20+ 8. we regularly ship extremely useful free tools for potential users, and each one ranks on Google and brings people into our products 9. every month we check which AI answers name our products and which don't Outrank's AI Visibility Checker asks ChatGPT, Claude, Gemini and Perplexity real buyer questions, then shows whether we got named, which competitors did, and which sources got cited we feed those sources back into Bazzly, Outrank and our internal social content all of this brings in new users, and then we run loops that turn them into champions who spread the word 10. we incentivize our users to share honest reviews of our products on Instagram, TikTok, X and LinkedIn, and on public review sites 11. we let our users become affiliates, and then we help them make better content about our products. in the last 30 days, ~$8K has been generated by Revid affiliates alone basically every SEO and AEO lesson I learn building Outrank and Bazzly goes straight into every product I run and the idea behind all of it is simple AI recommends the most mentioned product so we make sure every one of our products gets mentioned everywhere Google and AI look: all sorts of blogs, listicles, Reddit, review sites and social and this setup does the job extremely well, for every product we run
196
44
14
1,057
103,900
looking forward to this. curious how much of it is still classic SEO vs showing up in AI answers
4
Very few things in nature are true long-term exponentials. Most follow S-curves. I saw this when I worked at Facebook. What looked like clean exponential growth was actually a series of overlapping S-curves, each driven by technological breakthroughs and/or adoption waves. AI is probably the same. That doesn’t mean progress will have to slow down, but sustained exponential growth isn’t automatic either.
Updated version!
81
13
5
221
20,082
same with most a/b test wins. the first weeks look exponential, then it flattens to the real number
4
A customer service conversation I just had. We're all just meat proxies now.
34
3
1
52
7,832
ha. the support bot that can't do the one thing you asked is still the fastest way to lose a customer
3
To succeed in the short term, be too loud to ignore. To succeed in the long term, be too good to ignore.
223
388
28
4,420
95,718
the long term half is the hard one. loud gets the first sale, good gets the second
2
At @vercel we got to test this model before release. Best way to summarize it: astonishingly brilliant. One interesting takeaway: it's very clear that this new generation of models can take on unprecedented bodies of parallel work. To actually seize this intelligence, you need cloud sandboxes. Full stop. The model can delegate and orchestrate far more than a single computer can handle. Likely far more than you can even imagine it can handle. Your bottleneck is now ① your ambition and ② your parallel computing power. Btw, I learned from @cramforce you don't need to overcomplicate it. Install Vercel CLI, setup the Plugin and tell the agent to use /𝚘𝚏𝚏𝚕𝚘𝚊𝚍-𝚝𝚘-𝚜𝚊𝚗𝚍𝚋𝚘𝚡 skill. It'll provision and de-allocate cloud sandboxes as needed. You don't need a "multi-agent remote harness ralph orchestrator" 😆
Claude Fable 5.2 is live on AI Gateway. It improves on earlier Fable models in coding, reasoning, and vision, and can orchestrate or advise smaller models on harder problems. 𝚊𝚗𝚝𝚑𝚛𝚘𝚙𝚒𝚌/𝚌𝚕𝚊𝚞𝚍𝚎-𝚏𝚊𝚋𝚕𝚎-𝟻.𝟸
145
77
26
2,283
151,427
curious what changed most for your team, fewer review rounds or bigger tasks handed off in one go?
4
Btw the new meta is: - Fable 5.2 for planning - Opus 5.5 for execution Trust me, and thank me later.
242
231
44
6,748
246,850
same split works well here. a clean plan file in between is what keeps the execution model from drifting
2
Former Tesla president Jon McNeill on how you actually impress Elon: walk the Model X line until you find the real problem. The Falcon Wing doors had to be bolted on blind. "Have you ever thought about a jig so that they could just follow a pattern?" They built it overnight. The operating rules were just as blunt. "Do you think we could try to run the organization on three-sentence emails?" Here's my problem, here's what I think the solution is, here's my next step and what I need. And: "Give me bad news as fast as you know about it. Because bad news only gets worse with time." As Elon tells it, Steve Jobs "blew him off a ton," then one night at a sushi place in Menlo Park gave him the whole design philosophy in one line: "If the product is perfect, it will sell itself." That's how Tesla doubled every six to eight months with a marketing budget of precisely zero. And he doesn't sit on ideas. After asking how fast you type with your thumbs versus a keyboard, he had chief of staff Sam Teller pull the top brain-computer papers for the flight home. "By the next week he's meeting with some of the top minds in the world on this topic, and he doesn't let go until he starts Neuralink." 🎥 @TKPPodcast $TSLA @elonmusk
44
344
13
3,996
264,808
walking the floor works in ecom too. watching 10 real session recordings beats any dashboard meeting
36
Me: "I'll just make one product video." Also me: "Wait, I could turn this into a whole content account." Dreamina Marketing Studio generates the first video from product images and prompts. CapCut PC helps make it post-ready: AI Edit for visual fixes, AI Extend for longer product moments, and Multi-Track Editing to coordinate footage, text, and sound. One workflow, more content to share, and opportunities to explore affiliate marketing, brand collaborations, or paid video projects. Dreamina: bit.ly/4k0KTGe CapCut PC: tinyurl.com/2v23rn22 #CapCutPC #CapCutDidThat #Dreamina #Smallbusinessmarketing #capcuteditpilot
🤝 Paid partnership
89
116
393
46,324
too real. one video becomes a whole pipeline by friday
2
If you design with AI, this is all you need 1. sceneai.art — Beautiful AI Prompt Library, Made for Creators 2. land-book.com — landing page inspiration done right. 3. 60fps.design — mobile UI animations 4. recent.design — cool gallery for web design inspo 5. landdding.com — top notch landings save these for your next project
9
68
1
597
26,820
solid list. the tool matters less than having 3 references you actually like before you prompt
5
i've officially solved AI UGC end to end.. this is the most realistic AI video ad you'll ever see.. this entire video was 100% AI generated.. the biggest issue with realistic videos was always the training data of the models.. so i fixed that.. this is the frontier API for creating realistic images and videos.. nothing else in the industry even comes close this image model is the most realistic image model for UGC, drama ads, animation ads.. every possible use case this video model is the most realistic video model for ads and it's 10x cheaper than seedance 2.5 this is a preview of the miko API.. by far the highest quality way to create end to end ads for any use case this will literally 100x the speed and quality of your current workflows.. we're pushing out hundreds of high quality videos a day with agents + my API with this API you can make realistic AI ads for signifcantly less than anything out there. every movement and action in the video looks real because of how our image model is made and how the video model behaves paired with it the realism is 1:1 with a real video, and the videos work across organic AND ads at scale for any demographic or age with absolutely zero AI detection on top of that the miko API is the most reliable creative image / video API console in the industry: > every major image & video gen API at up to 90% off > seedance 2.5 at up to 80% off ($0.06/s) > seedance 2.5 1080p at 65% off > our own SOTA image & video models for the highest quality marketing material possible this is currently the frontier for making any type of marketing material, and it's going to change everything about how you create ads going forward..
50
14
4
254
23,494
looks good. real test is the ad account though, curious how it holds up on hold rate vs real creators
60
With the @runwayml plugin in Codex and ChatGPT, you can see how a character from your prototype moves, try a different opening shot for a trailer, or turn a product photo into a video ad. Anthony Maggio shares how Runway’s marketing team is using ChatGPT Image to produce more assets each week.
106
46
24
584
84,832
consistent characters across shots is what ad teams have been waiting for. that was the thing that made AI video unusable for brands
14
One post got 320 likes. The other got 80. Which one would you study if your goal was more link clicks? I gave Clover, the first ambient learning agent from @ParadigmStudy, a sample campaign report and asked it to turn that question into a mini lesson. I picked Post B because it drove 100 clicks vs. 64 for Post A. Then Clover showed me how to compare the two when one post reached a much bigger audience. For every 1,000 impressions, B got 50 clicks and A got 8. That’s the comparison I’d take into planning my next post. If AI universities eventually offer marketing credentials, I’d want them to show how someone made a decision like this. What was the goal, which number mattered, and what would they test next? You can try building a lesson around a skill you want to learn at paradigm.study.
🤝 Paid partnership
A perfect answer now costs nothing and proves nothing. So what is a university for? We're building the first AI-native university. A school of one. Introducing Paradigm.
19
106
193
29,033
depends on who liked it. 80 likes from buyers beats 320 from other creators
3
Apparently, every subtitle in my Black Friday ad picked its own font size. Love the independence. Hate the result. My budget-friendly workflow pairs Dreamina with CapCut PC. I generate product images and videos in Dreamina’s Marketing Studio, then import them into CapCut PC. AI Edit cleans up a selected background element; AI Extend leaves more time for the closing message. Then I tell EditPilot to make the captions consistent across my selected narration clips. One sentence handles the repeated formatting; follow-ups add a sticker and subtle sound effects. I can preview everything and adjust the text directly on the multitrack timeline. Try this pair for your next ad. Dreamina: bit.ly/4k0KTGe CapCut PC: tinyurl.com/2v23rn22 #CapCutPC #CapCutDidThat #Dreamina #Smallbusinessmarketing #capcuteditpilot
19
41
217
58,463
ha, at least it picked a personality. worth fixing though, most people watch those on mute
2
I had Fable 5.2 read: > 66 SEO/GEO articles > 48 skills > 14 GitHub repos If you need to bookmark 1 article, it's this one, not 66.
48
72
6
955
128,563
feeding it the sources first is the move. the output is only as good as what it read before it started
26
Breaking: It's been more than 24 hours since Claude Fable 5.2 launched. And People are already building mind-blowing things from just ONE prompt. 10 wild examples so far:
17
109
260
38,667
the best tests i've seen are the boring ones. give it a real messy task from your own week and see what it ships
126
Rewrote our entire portfolio tracking software from scratch with Grokbot + Claude today. Highly recommend. These tools have come a long way.
67
32
6
1,076
85,387
how long did the rewrite take? the part i'd worry about is the edge cases the old code handled quietly
106