AI-Powered Store Audit: $50K in 45 Days — Full Breakdown & Transcript

[00:01] experiment I wasn't sure was even possible. I wanted to see whether claw could audit a failing store, catch the leaks, and turn it profitable with only me managing it. And as you can probably guess from the title, it absolutely

[00:13] parts of my store, it caught the mistakes. It helped me fix them and in mistakes. It helped me fix them and in return made me $50,000 in 45 days. But money isn't even the real story here. What's interesting is the AI diagnosis

[00:27] loop that I built along the way. the one that I now run on every single store. for close to a decade. I've tried hundreds of tools, but none have changed directly solves one of the biggest barriers keeping people from becoming

[00:42] most valuepacked videos that I've ever made. So, y'all better lock in and let's go and jump straight into it. Now, before I get into the numbers and the AI show you what we're actually selling here so you have the full picture going

[00:55] in. Now, this product is called the Plump Tape, and we sell it right here this is is it's basically a mouth tape that you wear at night while you sleep. know that mouth tape has been blowing up in wellness for the last few years

[01:07] because it forces you to breathe through your nose, which is way better for sleep took it a step further. And this is where things get interesting. Our version isn't just a wellness sleep product. It's also paired as a beauty

[01:19] product because this tape is infused with collagen, hyaluronic acid, and vitamin E. So while you sleep, it's yes doing the functions of mouth tape, but it's also hydrating and plumping your lips at the exact same time. And that

[01:31] here because it means that we can market this product to two completely different audiences with the same listing, sleep wellness and beauty. Now, let me go and Surell storefront and right when you land on this page, you can see that we

[01:44] have the hero up top showing the product pouch with a 30 night mirror guarantee scroll down, you're going to see that you have the pricing tiers. Now, a single pack of this I was selling at $29.99, two pack at $49.99 with free

[01:57] the most popular option. And then the threeack bundle at $64.99, marked as have the rest of the funnel where you see like UGC creator videos, a science

[02:09] section explaining how the ingredients work, featured on logos for credibility. I even have this comparison table right here against competitors. And as we we have the FAQ answering every objection and even customer reviews with

[02:21] having all these parts is simple. By the time someone actually hits the bottom of they had is already answered. Which means the natural move from here is just that we did here, which is honestly one of the biggest things that you should

[02:35] offer this product as a subscription where the customer subscribes once, they the product automatically ships every month as well. And this is honestly the recurring revenue in drop shipping. So whenever you have the chance to

[02:49] with that picture in your head, let me go and show you what was happening on the store over the last 45 days because what I saw on the dashboard is what first place. So this is our Shopify dashboard right here. And I'm going to

[03:01] the net sales report. Net sales is basically just a total revenue, meaning store. And now what I want to do here is I want to compare two different time things are trending up or if they're trending down. Now the solid line that

[03:16] the dash line that you're seeing right here is the 6 weeks before that. So we have an easy comparison right here. Now the most recent 6 weeks has brought in around $4,282. The 6 weeks before that was $8,134.

[03:32] So we basically cut our revenue almost in half. Down 47% to be exact. But the when this drop started. So, let me go and zoom in into specific weeks to see when things went sideways. Now, the week of February 23rd through March 2nd, the

[03:45] store only did around $1,012, which is pretty solid knowing I was only spending the next week, March 2nd through the 9th, the store did $1,38. results from the week before. Now, it's still holding around like $1,000 a week.

[04:00] But, let me go ahead and jump ahead. Now March 16th through the 23rd, the store only did around $571. So you see we're trending a little bit downwards. And what we were currently doing in the previous weeks. And it didn't stop

[04:13] there. From April 20th through the 27th, the store only did $129. So I just went from going from a grand a week to basically like a hundred bucks. So somewhere around mid-March, the store literally fell off of a cliff. So the

[04:26] something called the conversion rate which is just the percentage of visitors just scrolling through and not buying anything at all. So in simpler terms if only three bought well that means that I have a 3% conversion rate. Now the

[04:40] industry standards for Shopify stores is around a 2% conversion rate. And you see my conversion rate is sitting at 3.92% up from 1.33% a month ago. Now that's a increase. And on paper that's going to look like a massive win. And this is the

[04:55] trap that a lot of new drop shippers fall right into. And now I want you to this. And this is the section about sessions. And a session is someone visiting your store. And for the current period, you see we have right around 102

[05:08] sessions. But for the previous period, we had around 300. Meaning that traffic actually bought in both of those periods? It was four, which is the same exact number in both. So after analyzing this, here's what really happened. The

[05:21] traffic crashed, the sales stayed flat, and the conversion rate just looked like a rocket ship because we're dividing four by 100 instead of four by 300. So, didn't get better at converting. It just lost 2/3 of its traffic. And this is the

[05:35] part most income report videos just completely skip over that conversion rate. Yes, it does look like a win, but it's actually just a symptom of the real store works fine. People are still buying when they show up, but there's

[05:48] something something cut my traffic by 2/3. And until I figure out what that is, well, the store just continued to keep bleeding. Now, normally trying to afternoon of clicking through dashboards, but this time I decided to

[06:01] do something different. I decided to see whether AI is capable enough to fix my store on its own. You see, one thing that so many people get wrong with AI and drop shipping is that everybody's using it to generate copy, generate ads,

[06:13] and maybe write product descriptions. And that's fine, but that's the surface level use case. You see, the thing that's actually going to move the needle Because the bottleneck for most people isn't creating stuff, but understanding

[06:26] stuff, like reading their own numbers, catching problems before they snowball. actually learned that Claude had came out with a new feature that lets you theory, they should give all of the data needed to audit my store, see where it's

[06:40] bottlenecks. Let me show you how all of Claw tab and I'm going to open up a brand new chat. Now, what I'm going to go a and do is open something called the connector directory, which is just

[06:52] tools or websites. So, now I'm going to click into Shopify. And then once it's on, Claude can now see my products, my orders, my customers, all my analytics. Basically, anything that I could see as the store owner. Once it's connected, I

[07:05] just run a normal prompt, something like, "Can you pull in the average order like, "Can you pull in the average order value from February 23rd to April 23rd?" order value or AOV is for short, it's just how much the average customer

[07:18] AOV just means you're making more profit you real quick what happens when I gave it that prompt. Now that Cloud is it's reaching into my Shopify store. It's pulling the data and then within

[07:30] full breakdown. Like average customer spend was $35.98 across the 45 days, strong early on at 49 to 55, then dropped to $26.99 on the slow days. And

[07:42] gave me the trend story, too, showing from late February into March, modest recovery midmon, then a bunch of zero basically giving me like a full analysis read on all of my data. So, now I'm just

[07:57] going to go ahead and follow up and say, what were the number of units sold in the last 45 days? And Cloud just pulled the answer. So, I had around 221 total units, the busiest day being February 28th with 12 units. I had a strong

[08:09] start, mid-March peak, and then April collapsed to one to three units a day. So, in literally about three prompts, I went from raw store data to a clear went from raw store data to a clear diagnosis. 221 units and a $36 AOV. And

[08:21] prompt pack in the description down below. So, you guys get direct access to the exact prompts that I use for every single store that I run without any back prompting. volume was concentrated in

[08:34] late February and then it gradually declined through March with a near flat just did, it took me what, like 2 minutes to complete. And now I'm sitting happening inside my store where the next thing I needed was analysis behind the

[08:48] I couldn't actually tell whether the problem was my store, was it my product, Now, if you guys don't know, I actually ended up finding this product on AutoDS consistently. And if I'm being real with you guys, it's honestly just because the

[09:02] doesn't just help me find a winning product, but it also helps me understand the entire market behind the product, too. And inside AutoDS, actually, one of is the ads spy tool. And this is easy because I always try telling you guys

[09:14] that content is king and data is queen. So, you want to see like what content is literally search and sort through the most engaging products on the market right now. So, what I did was I just pulled the product back up and I looked

[09:26] at everything that I kind of glossed over the first time around. So, here I'm actually being made for the product, what the comments were saying, the pain everything that I fully need to solidify

[09:38] this as my product that I'm about to pour way more money into. And what I telling. You see, the top performing like fancy. I mean, they were dead simple demos that were just hitting the

[09:50] And even when I came in here and I checked out the comments, I mean, you what I needed to hear, and that's that this solves a real problem. People genuinely want it, and they're already buying it from somebody else. And once I

[10:04] clicked. The product was solid, the market was solid, and the product market fit was there the whole time. So now I knew that the store was fine, and I knew that the product was fine, which meant one thing for me, and that's that the

[10:17] entirely. And honestly, knowing I don't have an answer right now, this is the part that I think is the most underused capability in the entire AI shift. I mean, everybody's so focused on using AI to generate stuff and they forget that

[10:29] AI is also the best analyst that you could ever have access to. It works for pattern matched across your data way faster than any human could possibly do it. So now I was in this position where I know what the data was actually

[10:43] volume problem, not a conversion problem. the traffic dropped, but the up, which means that the next question from here isn't, "How do I improve the store?" The real question is, what's happening in the ad account? And that is

[10:57] where this gets really interesting. So, the store didn't break. We know that the product wasn't the problem, leaving us with only one other place to look, which and show you exactly what that looked like, because I think that this part

[11:09] right here is going to be the most useful piece of this video for anybody Now, I just jumped into our meta ads manager, which is where I launched my different ad campaigns. Now, this top level view shows that I got 148 total

[11:21] campaigns running. Now, most of them are dead or paused at this point, but the story. Let me go ahead and start off with the one that actually worked. So, I this was basically like all my winning ads that are working the best at that

[11:34] time. And I want you to look at these numbers. I had $573 added to cards, $104 purchases, $3,989 spent, and $4,97 in sales. So, that's giving me a rorowaz of 1.23. And my break even here was 1.2.

[11:49] Now, if you don't know, rorowaz stands for return on ad spend. And it basically from every single dollar that you spent on ads. So, in this case, a 1.2 rorowaz meant that I broke even. A 1.3 means that for every $1 in ads, I was making

[12:03] just a little bit back. Basically, I was slightly profitable with this campaign. the store back in February through mid-March, which was that time where we want you to look at what happens after. April 23rd, I launched a brand new

[12:16] campaign. I spent $240 and I ended up having zero purchases. May 2nd, I threw out another test. I spent around $268. Still zero purchases. May 6th, $149 spent, zero purchases. May 9th, I still kept going. $86 spent. rorowaz of 69

[12:34] meaning that I was losing money here and then on May 14th 1307 spent rorowaz of41 which means I lost even more money. So here is what was happening. We had one the next 8 weeks just throwing creative tests at the wall trying to find the

[12:47] next winner. And here I burned about $2,700 and none of them hit and that's before I showed you what I did to fix this, let me show you what was inside the champion as well. So here you can see that there's 16 adsets in this one

[13:01] campaign. And an ad set is just a group of ads targeted at a specific audience. performing well here were like hair care us, Kylie Jenner fans, cosmetics, Drake fans, and a broad audience. And it's crazy because the biggest winner here

[13:15] was actually the hair care usually wild because this is a lip product and not a here, okay? We cannot predict what audience will work. We can only test decision. Now, if we actually go one more level down, you're going to see

[13:29] that we have the ads section with a total of 91 ads in this campaign. So, officially the full picture. The champion campaign died sometime in late March. And then 8 weeks of testing since then hasn't replaced it, and we have not

[13:42] the creative variety was there. I was launching a bunch of different ads. The hooks are there, but nothing has popped a rorowaz of over 1.0 yet. And even with running like five to 10 times higher than the champion baseline, which is

[13:55] basically meta telling me that your new creatives are not hitting. They're not them well. And here's the honest moment. I mean, if I just kept going on this and hoping something would hit, I would have burned through way more money

[14:09] wrong. Because what was actually wrong wasn't that the creative was bad. It was that I was so focused on hunting for the next winner that I missed how profitable the current winner still was. And this is the part where many people get stuck.

[14:22] They confuse activity with progress because running new ads or running new right. It feels like you're working. But you have to understand, I mean, if those is already working, then you're actually going backwards. Which brings us to the

[14:35] part that you've all been waiting for, which was the fix. So the first thing I campaign and I just looked at it again, but this time on a longer window, which was the full 5 weeks since we relaunched it. And here's what happened. So, let's

[14:48] workhorse ads. Inside of it, we have Talking1, Talking 14, Defined Hands, and Talking1, Talking 14, Defined Hands, and AD47. So, combined, I spent $3,814 in ads, and I brought in $4,743 in revenue with a blended rorowaz of

[15:03] 1.24. So, on a campaign that I thought was dead, I was actually still the most recent full week just to confirm. So, I'm going to go and pull up April 28th through May 5th. And as you see, Talking 14 had a rorowaz of 1.80,

[15:15] defined hands at 1.66, 66 talking one at 1.48 and add 47 at 1.11. All pretty much being a positive rorowaz. Now across the whole week we ended up doing 32 purchases on $1,555 in spend which brought $1,559

[15:31] in revenue given a blended rorowaz of 1.48. So compared to our break even of money here. So here's what was actually happening here. The champion campaign profitable the whole time. I was just so distracted testing new stuff that I just

[15:45] that our daily budget for this campaign is set to $200, which is where we were able to find the problem. Because $200 a day across four different adsets and 12 ads means that each ad is only getting about $30 to $50 a day. And at an

[15:59] average cost of $30 per purchase, each ad can only generate about one sale a day. So the campaign, it wasn't broken, it was just being starved. So the move is to push the budget up. And the first thing I did was push it to $325 where

[16:11] then we just hit save. The auction recommendation also says that there's is I can actually push that up to $500 a day. So, that's going to be two and a here, all we have to do is just click save and publish. And after making those

[16:25] from doing that. Because scaling the champion, it didn't just bring revenue back. It gave the algorithm real data to work with, real spin, real purchase kind of signal, it starts doing something extremely powerful. It starts

[16:39] telling you who your actual buyer is. And that information is exactly what led everything. So, I'm back in the ads manager and I want to show you exactly at full spin. Because inside this campaign, one specific creative

[16:53] completely separated itself from everything else and we call it V2. And without this ad, we don't get to $50,000. Full stop. Let me go and show you the front-end numbers first. So, V2 reached nearly 70,000 people with over

[17:05] 105,000 impressions. But the number that really matters is the clickthrough rate. And here you see that the overall clickthrough rate was right around 3.45% and the link clickthrough rate was around 3.09%. And to give you context, a

[17:18] 1% link clickthrough rate is considered decent in the space. And V2 was running definitely clicking and they were clicking because they wanted to. They saw the perceived value behind it. Now, here's the part that would have scared

[17:31] most people off. Now, the CPM on this ad was over $100 and that is extremely high. And if you're newer to this, a $100 CPM looks like a red flag. You see got to pull back. But here's what that number actually means. So, Meta was

[17:44] aggressively on this campaign because the algorithm recognized something that we couldn't see in the early data, and that's that this creative had the highest probability of generating a purchase. So, it kept pushing spin

[17:57] strong click-through rate and a strong conversion, that's not really a warning sign to be afraid of. It's meta telling you that it's officially found its buyer. And the conversion data confirms exactly that. As you see, we had around

[18:10] exactly that. As you see, we had around 93 add to carts at $11.80 each, 659 checkouts initiated. And the number that really matters the most here is $521 total purchases. Now, the total spend on V2 came out to $10,652.96,

[18:27] was $2045. And what did that 10,600 produce? And what did that 10,600 produce? $22,21467 rorowaz, which basically means for every single cent, every dollar that we put

[18:42] into this single creative, we are getting on average $2 back consistently at scale. Now, let me go to show you how dominant this actually was. So, here are every other creative that we ran alongside it. You're seeing V1, V3, V4,

[18:54] and all the rest. Now, combined all those ads spent around $1,200 in total. And together, they generated around 35 purchases where their average CPA was floating between 25 and $56, and their rorowaz was below 1.8. Profitable, but

[19:07] nowhere near as profitable as what you just saw before. So, that means that V2 generated 94% of the total purchases in this campaign. 95% of the total revenue, and every other ad we ran was essentially just background noise. So,

[19:19] the natural question is why? Why did V2 crush everything else so badly? Well, it wasn't luck, and it wasn't the algorithm randomly picking a favorite. The answer the other ads were product shots, graphic demonstrations, clean visuals of

[19:34] the tape, the packaging, the ingredients. Okay, decent ads. They were were structured in a way where they were trying to sell a product. But V2 did the a close-up of a human face. And that

[19:48] because people, they don't buy products. They buy outcomes and they buy relatability. They buy the version of themselves after the product works. And B2 led with that. Before someone even clicked the link, they saw the

[20:00] skin, the better lips, the more confidence, where the results, the real before the sale even started. That's the difference between a good and a great ad. A good ad explains what the product does. A great ad makes the customer feel

[20:15] it. And when you build that kind of creative on top of a proven campaign that's when the numbers start to look like this. That's the full story. The Champion campaign got the store back to

[20:28] stable. It turned the data that led us to V2. And V2, 10,600 in spend, $22,000 back, 521 purchases, 2.09 rorowaz is what pushed this store to $50,000 in 45

[20:40] days. Not because I found some magic winning ad out of nowhere. Because would have missed. I stop bleeding budget on dead tests and I scale what was already working and let the algorithm find the buyer. That's the

[20:52] loop. Connect claw to your data, run the diagnosis and stop starving your you a signal like V2, you fund it. You double down on it. And if you want to see everything, I mean show you exactly how I set this whole stack up from

[21:05] scratch. The products, the stores, the ad workflow, all of it. I made a full popping up on the screen right now. So, as always, thank you guys so much for watching. I'll see you in that next video.