The Problem

Let me help you. Ads have changed. The AI boom of 2020 is over. HAQM and Google still promote these AI campaigns as the most effective and they can be. The problem is not a lot of people understand the algorithms. HAQM and Google are not quick to tell you about these. You need to pull your data and find out where are you showing up and at what cost. Really dig the queries as well. Let me set an example for you real quick and show you what I am talking about.

Say you sell a thermos coffee mug. Sweet! We build an AI Campaign , omg look at all this exposure! All these clicks! Awesome, no. Not awesome , lets explore where this goes wrong!

Where AI goes wrong

thermos coffee mug grabs a keyword like Tumbler. Sweet, this is okay. It probably profits you. However the AI see's a thermos coffee mug as a cup. So now we are bidding on cups, okay, maybe someone will buy, it also see's it as a sippy cup , it's not a sippy cup, but here it is popping up for every mom and baby shower person to look at and they are clicking on it. When they realize it's a coffee mug they click out, who wants to buy a baby a coffee mug it wouldn't work right?

So lets go deeper, we are now bidding cups but say your thermos coffee mug is Stainless Steel. Cool. AI doesn't care it put more weight on the cup aspect. So when someone queries "plastic cup for house" your stainless steel mug appears! Sweet! No, not sweet. The reason is you will not meet the expected price range or the functionality that query is targeting.

The problem with AI is it uses broad based targeting. The objective is to put you in front of everyone it thinks you should be in front of. This includes Plastic, Aluminum, wooden cups, then we add tumbler to that equation and variations and suddenly we are showing up for a lot of stuff we aren't. This hurts because those terms are more frequently searched so even if 90% of them are not looking at your cup the 10% that do and click are eating your ad budget. Meanwhile the chance of purchase is extremely low on those queries because you aren't as relevant and that means you shouldn't show as much but !! You wanted more sales so you raised the bid from .87 to 250% for featured.

Now the 8 dollars for a 10 pack of plastic cups is bidding with your 19.99 single stainless steel coffee mug/tumbler. However those cups are bidding 0.30% with 150% for feature. So even though you are less relevant the other side of that equation is the bid! You are such a profitable bid that you are showing near the top despite not being what was actually queried and I know you all see this in your own results searching HAQM. When you look for something cheap and that expensive , upgraded one just pops on the screen and we can't help but click. When you're shopping cars there is a reason they place the most expensive and nicest vehicles on the front page, there is a reason despite using filters some of those more expensive cars still pop in you search results. 75% of us are going to find the cheaper car we wanted but 25% of us are suckers, we are trying to figure out how to fit that leather with the upgraded trim and wheel package an free oil changes for a year into our budget. How can we make this car work. We totally forget about the Kia Soul we could afford and now we have a Telluride fully loaded and we are hooked into something we should never do.

HAQM and Google use this same concept but it's a concept that doesn't work on a broad scale. The days of bidding on as many keywords as possible are over. People are not in spend more money mode. They are being specific about what they want. Now if your plastic cup was 20% more but had cool things like a lid and straw sure, but your stainless steel coffee mug is 300% more and you get just one! Not 10.

So how do we stop this? You should run automated campaigns product only. Do not fall for the brand or off amazon ads they are a waste. Brand works if you have an established brand it doesn't work for anyone else. Off HAQM ads get so many bot clicks it's unreal. It won't deliver.

How to fix this!

Auto Campaign (AI)

Negative Phrases - Plastic cup, plastic coffee cup, sippy cup, (maybe go further), Cup, wooden cup, cheap cup, household cup, aluminum cup, plastic tumbler, aluminum tumbler etc...

Bidding - If you pay 5 dollars for the cup landed and all, you need to figure out a few things. What's the target profit margin, say 30% , no what does amazon take? 15% , now what does it cost to ship? add 6-8 , returns and fraud? add 2-8% depending on your category. Add budget? Should be around 15% max. Basically I tell people the way to stay in profit is to take the COGS and multiply it by at least 1.7 and that's your selling price before shipping. In this case your base price is 8.5 , giving you 7.22 after amazon fee 7.22-5 = 2.22 per unit leaving 26% minimal margin before returns and fraud. However I suggest going 1.85 giving a selling price of 9.25 instead. This gives you more room. 7.86 - 5 = 2.86 2.86/9.25*100 = 30.9% margin. This is healthier but you goal should be 40% average or higher because this is before we pay ourselves and everyone else and rent and all of that. So now we do ourselves a favor. We add the ad spend to the actual price. so if its 9.25 and we are targeting 15% we end up with 10.64. We add shipping in this case I'll add 8 (remember AMZ take 15% of this too) , 18.64 , this is the min price we can sell for. Now if we find everyone else is at 17.50 , well we have a few options. We can use less of our budget on ad spend, we can lower our margins slightly (but not much room to move), or we can find a better shipping solution.

One we establish our budget we stick to it. So 15% if your daily sells are 500 dollars you are spend max is 75 dollars a day.

So now we have a budget, 75 a day or about 2250 a month. The math now maths! Sweet!

Now we gotta decide how to tinker and spend that 2250 to drive our sales higher than the 500 a day.

Suggested Bid - Do not use HAQM's suggested bid but you do want to consider the range. If it suggest for example 1.30 bid but the range is 0.75 - 2.95 you want to start low. I would set my sights around 0.55 and move from there. Then for the featured you can meet that range. Say 50% increase , now it's 0.80 which is in that range above.

From there you can tinker up or even down. If you are getting a lot sales drop in a few pennies.

Here's the thing about scaling though. You want to be profitable no matter what. So pushing to get higher numbers just isn't it. Growing sales as fast as possible to reach higher rankings is a Strat only larger companies can use. We lost 60 grand doing this but 60k to my company was nothing. However, our new products I don't do this. I scale them up slowly over time and peel back ad spend as we increase organics or we'll increase ad spend to promote our rank , the overall goal here believe it or not is that organic.

I hope this helps someone. It's not something Youtube talks about or HAQM or anyone it seems. Ad companies just spend as much as possible on broad results. You want a target ROA. You target should be in the profitable range. So if you are spending 15% of your sale on ads as we show above your ROA target should be around 6.6 , assuming all your sales come from ads. Now if half come from organic that's great news. Because now that ad spend we built in from the organic is helping the paid ads because it wasn't spent. So overall we need an ROA of 3.2 , which is a lot easier to achieve at first.

However a target ROA of 6 isn't unreasonable despite what people here may say. Believe it or not there are people who sell products not at the cheapest price and get ROA's over 10 based on SEO, Organics and pure bullying of the ad system but these products requires a lot of content, a lot of targeting, a lot of A+ and a lot of well spent Ad spend.

The math we started with is the most important part of everything you will ever spend on HAQM. It has to math. If creates your boundaries. If you follow it you can never lose money.

My advice to you is start pulling all of the data. You have it in the ad section, in the reports etc... pull it. Look at it. Get all your cogs , find the average shipping cost of each items , etc. Then start doing the math. Figure out the number.

Then adjust the ads as i showed above. You can use your own data to find specific exact match queries that are hurting you budget instead of broad. Remove those irrelevant query ad spends. Focus down on what your product is and compete with others.

FBA vs FBM

Now if you are FBM or FBA there are other things we can do. For example the shipping we can offer free shipping FBM but if you can budget a lower shipping right you can offer better pricing in bulk which will lure more business customers. This is what I do mostly. End Users buy my FBA free shipping listings for say 10.99 , while my FBA listing is 6.99 plus 5.99 for shipping , but if you buys 10 now I'm 6.99 plus 12.99 for shipping about 80 something where my FBA would cost 100+. You can sell a product FBA and FBM by listing different conditions and you can pace them correctly to compete with each other based on bulk vs non bulk purchases. This is a whole other aspect but there's lots you can do. This involves building weight based shipping templates and doing calculations to make sure you are in the range. You may lose or gain money on shipping but it should all wash out. I am in Houston so I do this calculation to Seattle meaning worse case scenario , I usually cover the cost. There are some parts of lowly populated areas we do lose some money shipping to but that's okay because there are areas close to me and on the East Coast I actually make a bit off shipping. It breaks about even in the end but that took a long time for me to tinker down.

I hope this helps someone it's advice you don't usually find out there. We hired an ad team for our company. I had consistent 7.2 ROA's at 200k a month, they increased us to 350k a month over 6 months but our ROA declined to 2.9 and we we selling the majority on ads. I had to show my boss we were making 86k a month on my 200k vs 32k a month on their 350k before it finally snapped for her. Scaling is done slowly and I have worked harder to increase us to 300k a month while only dropping the ROA's from 7.2 to 6.9. I am not moving forward until I find a way to average 7 again over a 90 day period. This is how everyone should approach ad spend. ROA or ROI all that matter is that the math, math's. The rest is tinkering.

Manual Ads - This is like automated ads but far more difficult. I would love to drop a guide on this but there isn't enough room here for me to do so. However, well tinkered Manual ads as far as ROA and ROI's go will exceed AI ads by a lot. I have some that do 17 ROA's. The problem is the sales will be much lower and that's okay! You can use that extra money to pay for those higher ROA ads that deliver more buys.

In Conclusion

Narrow your budget and figure out the math.

Remove Sponsored Display - until you start getting decent profit these are a waste. Bot farms and placements mean people click this or bots do and they have 0 intention of buying form you.

Remove Brand ads - also a waste. You can also try bidding extremely low for these like if you are bidding 0.8 on Sponsored Products then bid 0.25 on these. Spend time and resources elsewhere (A+ , marketing etc) building your brand before you start promoting something no one knows... Otherwise keep the spend very low.

Use the data HAQM and Google Ads provide you to find and target Negative Keywords as Exact and Phrase. Remember to only use exact for exact terms and phrases for general terms in a query.

Do not use HAQM Bid Suggestion- Find the price yourself over time with trial and error , start low and build up.

Set a budget cap - Ours was 2250 per month in this example, we do not exceed that. If our sales are poor we just focus on using that budget and tinkering to make it work.

Ads take work - A lot of sellers spend less than an hour a month on their ads. This isn't enough. The ad company we hired spent about 4 hours. I can see how they would help some sellers but I spend about 12-20 hours a month on this myself. You probably don't need to spend that much time as our catalog is over 5000 products, but I would set aside 6 hours a month to do this. Dont focus best practices, focus what the data shows you works for you and your product.