Help with PPC

I’ll kick this one off - Handmade seller, very small by AMZ standards of course, 200 products +/-. PPC is a mystery to me, but I’m trying to figure it out (or if it’s even worth it). I have keywords & SEO pretty nailed down.

  • Last year (to make it easier to follow) I set up a separate campaign for each of my product types, then a separate ad group for each product - I have variations so just included primarily the lowest price one.

Let it roll, ran PPC reports, added lots of negative keywords that don’t apply to me.

  • In November, got an offer from AMZ to set up campaigns for me - They set up two, one with a low PPC that has most of my products, and another that only has around 10 with a higher PPC. I went in and added the neg keywords.

My questions:

  1. I ran a report yesterday and found 800+ clicks (and only 1 sale) from an ASIN that’s used in the same room as mine, but nothing like mine. So I went in and added that ASIN to ‘Negative Product Targeting’ - I assume that’s the correct thing to do? And now I figure Amazon will just move it to another non-related product so it’ll be like daily whack-a-mole.

  2. Ran a report just for yesterday, and my PPC costs were much higher than my PPC - I checked the campaign settings and it’s dynamic up and down. So that explains it, but I’m not sure if that’s best…or an Amazon money grab. Leave or change?

Thank you for your patience & any advice!


I never do dynamic up and down. Only down.
Amazon decides the cost, counts the clicks, and decides to whom they show the ads. It’s horrible.

I’m against anything where Amazon takes the initiative. So if Amazon suggest the campaign/shipping automation/ customer service, I immediately say no thanks (when they do it over the phone, sometimes they pester with calls to offer the services,I’m quite explicit about my negative, so as to get black listed from the calls)

I don’t have a lot of experience with negative keywords. Usually I run my campaigns with very few key words and what I do is run them for 2-3 weeks, check the acos/roas (they are the same thing but different sides of the fraction) and basically (very loose rule of thumb) for words/keywords

  1. if it does not get clicks, I leave it as there’s not enough data to make a decision
  2. if it has 5-10 clicks and no sales I kill it.
  3. if it has 5-10 clicks and just 1 sale, I let it live for a bit longer as there’s not enough data, regardless of the acos/roas
  4. if it has spent +20% of the budget and the acos is +60% (acos ), I kill it.
  5. if it has spent +40% of the budget and the acos is +50%, I kill it
  6. if it has spent +50% of the budget and the acos is +60% I kill it.

The idea of course is to work to lower the ACOS as much as possible. My best campaign has an ACOS is the single digit. I often wonder if I should just kill everything else and put all my budget in it, but it also feels like betting all on red…

I start my campaigns with very low budgets ($35-50/day maybe) So I’m forcing myself to work with limited data. But I just don’t like squandering cash. If after 15 days or some $500 I’m spending all the cash and not getting sales, let alone at least matching my $500, I reduce the budget and check for another 15 days. If after that is still no go, I kill the campaign all together.

With all this said, I’m conservative to a fault with money. Work too hard to give it to Amazon. And currently I’m getting slammed by a competitor who is positioning themselves with PPC in a way I’m not able to touch them. Not sure if they got good at it, or simple are burning through piles of money.

@ASV_Vites is well versed in this part of the mess. Hopefully they can give better insight


Our overarching rule for advertising is that for every $1 I put in, I should get $4 back in sales.

Now, advertising seems to have two kinds of value: sales and impressions.

I have kept campaigns alive that don’t generate a lot of sales just because the number of impressions was astronomical.

In the PPC interface, the sales you’re seeing are essentially impulse purchases - they see your item advertised and they buy it right then.

I have a theory that this misses a lot of sales. We have had items that were clocking decent sales ranks evaporate entirely when not advertising, even though the advertising campaign didn’t appear to be doing that great.

Our campaign layout, in a perfect world, would be to have an auto campaign with dynamic up/down bidding set lower than our best performing keyword. Dynamic bidding will bid up to double what you put in.

Alongside that we would keep a manual campaign with a slightly higher budget where we would steal keywords from our automatic campaign. This campaign would bid down only.

We take keywords by negative targeting them in the auto campaign and then putting them in the primary campaign.


On mobile here so I can’t write as much as the previous response. But I agree that we never use dynamic. I do not trust Amazon’s determination of when a click is more likely to result in a sale.

On a side note. You may want to try straight product targeting competitor Asins instead. You won’t have the wack a mole that you are describing. Just pick where you want to show.


About to sit and absorb all of your excellent advice - can’t thank you enough for taking the time to write it up, very very helpful, and I have a lot of work to do!

I’m usually a “thanks but no thanks” with Amazon’s advice, but figured it couldn’t hurt since I can always turn it off. But it has been a costly experience, sadly. I think I’ll go back to my campaigns that are more detailed so I can really watch it like you indicated in your post, Tally.

Re: Negative keywords: I sell art, but not canvas (which would have to be drop-shipped). And I kept getting bad feedback from people who assumed they were getting canvas (listings are clear, but people don’t read or look at infographics, as we all know). So I added all the things I’m NOT into the negs. That has helped a bit.

Haven’t tried a manual campaign yet, nor product targeting - adding these to the to-do list to try. Thank you for mentioning those.


The thing about auto campaigns is that they’re based on Amazon’s search algorithms. So they know what people are typing.

I have found that it does a fair job of finding functional keywords a lot of the time, but it needs babysitting. It will do dumb things.

For example, we have a TENS unit that we advertise that has a good campaign. Substitutions it shows on are our competitors and we see results from searches for pain relief and EMS and stim and all that.

We moved some of those awesome words into a manual campaign and cranked the budget up.

We also saw a lot of… Junk words. The letter M is a very expensive keyword. No one knows why. We saw butt plug, pregnancy simulator for men, prostate stimulator… You get the drift. It actually caused us some trouble, because it got our item flagged as adult content. So we had to neg out all that junk.

Sometimes it would grab weird substitutions, too - like advertising on a sofa and an espresso machine.

I’m not sure what all of that’s about.

We also used some of this information to change our listing because we wanted to just show up in searches that were advertising well.

So I guess my advice would be to use an auto for a little just to see if people are searching things you didn’t think of, but don’t let it run wild.


Thank you and yes to all of that - I had auto campaigns (now rewritten) for a few years that helped me hone my listing keywords & also find the dumb negative keywords I needed to add. The 800+ clicks (ugh…I feel so stupid for missing that) were from a ‘galvanized shelf.’ Seriously? Into the negative products it goes.

As an aside: Just saw some women in my town FB group saying how amazing TENS units are for menstrual pain. I know medical claims are tricky - but thought I’d mention it! (we love ours).

1 Like

Our core business is actually prescription TENS and clinical electrotherapy - we have a government contract that basically pays all of our business expenses. They’re good machines that really do work well, but we always take care not to oversell them.


There’s a few things missing that will help with necessary changes. I will speak in broad generalizations as foundational things tend to be better adopted and incorporated into an SOP.

  1. As Tally pointed down only is the way to go especially/and/ if you’re using software for bid optimization. However, manual campaigns that are manually adjusted will depend on your aggressiveness on a keyword.

  2. The targets others mention work in a strict mathematical sense, but may or may not coincide with the goals of running campaigns and advertising. Essentially, any ppc expert will present two things when it comes to advertising and this is also consistent with amazon’s own presentation - The Sales Funnel and the Amazon Fly Wheel - you can google both if not familiar with these.

  3. The more you feed these two concepts the better you rank/convert - if your products (offer) conversions are better or comparable with competitors.

  4. You said you have more than 200 products but how are they set up? My guess is that many are variations?

  5. Are you a brand that people seek out?

  6. If 4 and 5 are true then I as a best practice I would make two suggestions: For branded keywords - let’s say your brand is “ACME” and your branded search term is “ACME doohickey” - then you want to outcompete everyone with the highest acos on those terms - eventually the conversions will bring the acos down tremendously if you convert for those terms - also it would only make sense if your listings are set up with your brand name clearly visible in the title.

  7. Next suggestion is that you want to advertise the highest converting variation only and not all other variations - those are extra conversions which will then decrease your TACOS. - this is more of a high commodity listing trick - to boost conversion rate and decrease TACOS.

  8. As far as negative matching - normally - again if using software it’s easy to just lower bid rather than negative match UNLESS the ACOS is 100%+ or conversions are 0 over the course of 6 months - but this will also depend on how many impressions and clicks you got for that particular product, kw.

  9. The core of PPC is isolation - they best converting kw should be in their own campaigns and isolated to allow full control - these are all exact match types.

  10. A lot of these answers are dynamic depending on who you are, your product type, your niche, your brand, your revenue/cashflow goals etc - so very hard to give you specific targets.

  11. Remember: You are always competing with amazon in a sense - they always want to push you spend more and you want to only spend if the ROI is worth it. By design, you have a lot of amazon ppc managers who have a rudimentary understanding of PPC and then defer to another manager - but most have not been around to determine how best to bring costs down and increase rank. They are experts in new ad types, betas and pushing you to spend more.

If you have specific questions; glad to help.



Yes(ish) to both - 200 parents, all with variations.

Yes/no to the 2nd: They typically search for my brand on Google (which gives my website first), but not specifically on Amazon.

No software, I’m handmade and very small, so just me. Definitely getting the gist I need to set up a manual campaign (or 5) with all of the great advice you’ve all given.

I can’t tell you how hard it was to find this out - I had no idea whether I should advertise them ALL or just the most popular. I decided they would compete and just did most popular, so that’s good. I use that same logic to decide what to send to FBA.

I honestly use negative matching because of grumpy customers. If they search for ‘canvas’ and Amazon shows them my listing, they assume it’s canvas even though canvas isn’t mentioned or shown anywhere - it says print, unframed print, print only, not framed, etc. in the titles, bullets and A+. Crazy to me, but there it is. I had enough ‘Not as Described’ returns to know I had to fix that - and fast.

Reading and taking notes - thank you so much for the feedback!


Your response added clarity to your niche. So additionally:

  1. When you said 200 parents my mind was leaning towards high volume - just what I’m used to :slight_smile:

  2. The specifics of customer error in your niche is very hard to address. In my niche cross errors actually lead to conversions but in your case the confusion between canvas and print makes all the difference for “not as described” and such which is a nightmare - so now I understand and you are absolutely correct - here it isn’t simply an issue of conversion but also “customer error” mitigation which affects your seller performance - I would, at will, make negative everything that affects you thus.

My pleasure!

1 Like