Michael Hartman of IncPlan has helped set up more than 1,000 US Amazon companies and other online businesses. His Summit presentation covered the importance of PPC, and how to structure and optimize campaigns.
Presentation title: Supercharge Your Amazon PPC
Speaker: Michael Hartman, European Director, IncPlan (USA)
For more useful info about selling online, check out Global Sources Summit, a conference for online & Amazon sellers held every April and October in Hong Kong. Learn more: www.globalsources.com/summit
Supercharge Your Amazon PPC (Michael Hartman) – Global Sources Summit
So that’s a little bit of background about me. So I have 10 Pro-tips for successful sponsored ads. How many people are running sponsored ads here? Okay, quite a few. How many people are running AMS? Okay, not so many. Alright.
So ‘Amazon PPC Advertising’. We have Sponsored Products, we’re going to go over Campaign Structure, we’re going to go over PPC which is Plans, Proactivity and Customer behavior analysis, optimization, and then I’ll give you some tips and I hope that you actually get some value out of these tips. I’m going to give you some, but not all.
Amazon PPC Advertising Sponsored Products. So most you guys know if you have sponsored ads. These are the positions of your ads. Let’s see. <Oops, nope. Wrong one. There we go.> So on mobile, you have here at the top. On the desktop search, you have at the top and then along the sides. These are just some generic ads. Consider using Sponsored Ads to increase traffic to your product listing, which inevitably will increase your sales, ultimately increases your Amazon Best Seller Rank, and increases organic visibility and organic sales. One thing people always mistake is like PPC is disabled. It’s not. It will never be more than 25% of your sales. Never, right? But it will help boost your organic ranking or help fight off competition. So those are the things that you can do with PPC. It’s not going to give you like a save-all-for-your-product, it’s not going to fix something that’s broken. So sponsored ads dominate the Amazon search, think that only 30% of Amazon shoppers make it past the first page. Seventy-five/seventy percent of Amazon shoppers only purchase products from the first page. And this is a way to get on the first page through the sponsored ads.
Campaign Structure. Keep it simple. Create a campaign for each product. If you have a large inventory, group similar products into the same campaign and split them into different ad groups. Name your campaign so you know what the products are in it. A lot of time I see people I’ll go into campaigns and they have no idea they’ll put bedsheets under something that’s completely different. So the name of the product/thing will be like sports and outdoors and they’ll have bedsheets in there. That’s not obvious, right? Start with automatic campaigns until you get enough relevant data. So like two to three weeks on average depending on the demand of what you’re doing. Automated targeting system serves your ads to all relevant customer searches based on your product information. Also test out different bids. If you’re not getting impressions, obviously you need to increase your bids or you need to look at your back-end on your listing. Because sometimes your listing on the back-end it’ll affect your automatic targeting. Sometimes you’ll run into errors there.
Here’s an example of an *automatic versus manual campaign. So you have one auto and one manual. You always run in auto and manual at the same time. You’re going to take the best performing keywords out of the auto and put them into the manual and adjust the bids and you’re going to take the worst ones and add them as negatives. So you have ad groups or variations, large and small.
So we have *Plans, Proactivity, Customer behaviour analysis. Strategies influence goals and optimization (target ACoS). So every customer that comes to me they have like in they want a goal. So they want to be on the first page, they want to rank or they have a lot of competition coming after their listing. So there’s different things that you can do. You can have a target ACoS of a lot of people when they want to write they want to do breakeven. So that way they’re not losing any money, but it’s helping increase their organic sales. Because the more that they’re getting conversions on those keywords from PPC is driving the organic up so it helps with sales velocity. Sales velocity is key when you’re ranking on Amazon. Analyse everything from your data—your customer search queries, your impressions, your clicks, your cost per click, value of sales, your ACoS. Analyse reports to get new keywords. And one thing that a lot of people forget is conversion rates. If you don’t have a decent conversion rate, don’t even advertise, right? Decent conversion rate I have to say is over 10% that’s on the low side. A good conversion rate 25-30 percent, so that’s a real good conversion rate. Because if you want to maintain that first page ranking especially in the US market, you have to have something between 20% and higher on most products.
Create a strategy for each product. Choose a product, start an auto campaign, start it with a small daily budget and a small bid, analyse the results after one week, and if you don’t have enough impressions, adjust the bid up, right? So that’s one thing. Let campaigns run 2-3 weeks, then you can get in to create manual campaigns. But remember, always leave the orders going because you would be surprised that the orders will always pick up new language, new trends, new other things that just didn’t happen before.
Be proactive. Analyse search terms, add negative keywords. When you’re adding in negative keywords, only do exact. Do not do phrase, right? One example I can give you is let’s say, we have ‘kids’ red bicycle’ doesn’t convert. But ‘men’s red bicycle’ converts really well. Doesn’t really make any kind of sense but that’s just what it is. So make sure when you’re doing the negatives, you only use exact. Don’t use phrase. Don’t use broad. Always exact, okay? Exclude the irrelevant searches that are spending your money but don’t generate sales. Add the keywords that generate sales into your manual campaign and you can normally jack the bid on those. “Not everything that can be counted counts and not everything that counts can be counted”, said Albert Einstein.
*Customer behaviour analysis
Customer behaviors can’t be counted but it counts a lot. There are many types of customers—the ones who search exactly what they’re looking for they are looking for, they are looking for a specific size or color; customers that search for some product category and they are attracted by your product and buy it. So we can consider all of these factors when we’re doing advertising especially when you’re looking for keywords or adding new keywords into your manual campaigns. You can find in your data some new keywords that are not eligible for your product description (but they might be eligible for your product category) and they generated at least one order. Add those keywords in your manual campaigns and test them for one week. If they don’t convert, pause them before you spend too much money. But come back every so often and try them again because I’ve seen this happen a lot where one month the keyword will convert like crazy, the next month it won’t convert, and then four months later it starts converting again. And it’s just language. Language is always changing. That’s what I attributed that to. How people talk is always changing. Always analyse the recent data. The list containing the most relevant keywords may change from month to month. So one month it might be one word, another month it might be another word. And it’s actually quite interesting.
*Advertising Reports. How many people know how to build a pivot table? Okay. These are very important. You have to use your pivot tables if you want to do anything but the most accurate thing out there and stay out of the beta mode because it’s garbage.
Optimization. Be patient and do not optimize too early. Adjust keywords bids. Pause keywords, add new keywords. Adjust campaign budgets. Pause ads. Change keywords match type. And negatives are very, very important, but only on the exact. Because otherwise, they can eat up all your budget.
These are some tips that we have for when we start looking at a client and we have to decide whether or not we can start advertising or we can’t start advertising for them. So the first thing is the title and description of your product. If you don’t have a good title on description, this will affect your PPC. This will affect everything in the campaign. Everything’s tied together when it comes to PPC and organic. Add high quality images. If you don’t have good pictures, go get new pictures. Look at your bullets. Make sure that they’re good and they’re solid. So like I was saying, if you don’t have decent conversion rates, don’t advertise. You’re just wasting your own money, right? Or if you have really low stars, don’t advertise. You’re wasting your own money. Get a different product. Add relevant keywords to your product using all match types—broad, phrase, and exact. Use other advertising channels to bring traffic on your product listing. How many people are using Facebook ads directly to Amazon? How many people are using the Twitter link from Amazon for Facebook ads? One guy. That will help you rank your system pretty quickly. If you didn’t know that, it makes Amazon’s algorithm at least right now look like it’s a viral product. So people are tweeting about it. So you’ve run Facebook ads using the Twitter link. That’s a little trick. Structure your campaigns with only one SKU per ad group, right? Run both automatic and manual campaigns for your product as both have their own benefits which we went over of the automatics always generating new keywords that you can add into your manual. Use Bid+. Amazon says that Bid+ gives a better chance for your Sponsored Products to rank for searches. But in our practice, Bid+ is absolutely necessary to get to the top of page one, right now. Even if there are no other Bid+ bids. Basically, if you want your Sponsored Products to rank at the top of the list, you need to be using Bid+. Bid+ only works for manual campaigns, and it’s hit-or-miss per product. That’s what I’m going to tell you. Some products it works very well; other products it doesn’t work. And it doesn’t seem to make sense. Test singular automatic campaigns. This is very, very good piece of advice. Run an automatic campaign with all of your SKUs in one ad group with a small budget of 5 cents per click and it will always allow you to find the most relevant keywords that you’ll have the lowest ACoS on this at any other campaign that you do. Especially the more products you have. So if you’re rockin’ 50 products/100 products, I’ve seen it go down as low as like 0.2% and it just stays there. I mean, it’s great. So definitely try this. It’ll help you pull a lot of long tails out which then you can like insert into your back-end. I’ll help you with some organic there, too. Run multiple manual campaigns for the same product varying bids and keywords. Or, so you can use phrase in one and exacting the other. This way, when you’re doing the advertising, you’ll have the top position, the side position, and the bottom position. So it gives you three different chances to make a sale. It’s really key for ranking. If you’re trying to rank, you can run multiple manual campaigns, the ACoS won’t be so great, it’ll be probably 30-35% but it will help you rank for those keywords.
And that’s it. You guys have questions? Because I figure that there will probably be some questions. Yes?
<question; audio low> You can run multiple campaigns with multiple different bids. And it works. <question; audio low> No. Each campaign has different bid structures. This way, like one campaign that I set up recently, the entire budget was like $30,000 a day, right? I’m not spending $30,000 a day but I’ve ran a whole bunch of campaigns on small bids—10, 15, 25, 30 cents. All exact and phrase match and it actually on most of the pages, it was it took up two ad positions. And then if you get in the 2nd/3rd pages, you’re going after the other like 30% who are buying on the first page, it was insanely really good conversion rates and really low ACoS altogether. Like single digits for all of them.
<question; audio low> If they’re having like bad performance, then I just get rid of them, right? And I start over and sometimes like I could tell you like last summer you used to be able to take your AdSense of your competitors and put it in. It would work for two days and we just stopped. Amazon adjusted the algorithm back in December. That stopped working. So they’re always adjusting the algorithm and you always have to be on top of that.
<question; audio low> Amazon bases it on the history of the campaign most of the time unless they had a decent ACoS or they were achieving something that they wanted. We start all brand new campaigns. It depends also I mean if they have a decent ACoS they have a lot of history, then maybe we’ll leave that one on and we’ll create a competing campaign underneath that and see which one does better.
When you use Facebook Ads, how do you measure the success? I even number of conversions that happen because of the Facebook Ads? There’s a few different ways to do that. You can do claim codes, right? But the claim codes everybody will get like 50 and only one or two will convert. You can reverse engineer it, as we say, with like high-low profit and look at your conversion rates before and after. But there’s nothing 100 percent. But we mainly are looking at if we’re steady on a product that we’re not growing anywhere and I start pumping Facebook ads directly to the listing. Then I’m looking at—is it moving? Is that listing moving? Is it not with moving? So that’s it. I mean, there’s no exact science on that one.
<question; audio low> Depends on what your profit margin is and what you want. If you want to rank, normally, I say go breakeven and that way you’re not going to make any money there but 80 percent of your income is normally coming from organic, anyway. But you’re able to shoot up and fight off competition pretty hard. If you just want to optimize, you’re already on the first position, you just want to fight off competition, then you can target on AMS, your own buy box. You can do like long tails and the main keywords and just go for that and get the ACoS really low. It depends on what you want. Also, never run out of stock. So always, don’t pause your ads but reduce the budgets drastically to like $1 / $2 per day, right? Don’t pause the ads but drop the budgets $1 / $2 per day if you’re going to run out of stock. Don’t run out of stock like that’s the one thing I’ll tell you and don’t pause the ads like unless you want to kill that ad set. Because if you’ve restarted, Amazon will penalize you right now. I don’t know. I mean, but the ACoS just shoots up for like a week and then it drops back down. This is, it is what it is.
(Are you) Using the software to help you monitor the campaigns? No, we threw all that in-house and we have some in-house software that we use but we don’t use any you know PPC scope or anything like that.
With the pausing of the campaigns, that’s what I typically do. But do you think that you can still come back or you just starting a new campaign better than or…? It depends on what you’re doing and what you’re selling. Most of time if I’m pausing a campaign, one, I’m like throwing a Hail Mary because I’m almost out of inventory and things are going to go really bad. Or two, it’s just not converting and I just pause it but I want to save that data for a later date. If that makes sense.
Other than simple Excel pivot tables of the software you use for analysis or data analysis? Other data analysis that I use? We have some in-house tools but I’m also a big fan of Helium 10 and Keyword Inspector so you can do reverse PPC and look at your competitors’ campaigns, see which keywords they’re going after. What’s the first one? Helium 10. I look at that a lot, too.
Any more questions for Michael?
Where was that Twitter link you were talking about? If you go to amazon.com on a product page in the upper right hand side you’ll see a Twitter logo. You click that, it’ll say ‘share to Twitter’. You copy that link over and then you put that into your Facebook ad. So you’re using that as direct? So basically to Amazon’s algorithm it looks like you’re getting a lot of tweets. Tweets have let’s say more ranking purposes than Facebook. That’s basically why we use that link.
Any more questions?
What do you consider a statistically significant amount of impressions, clicks, click-through rate to terminal, pause a keyword or turn it off? It depends per keyword and per product. Depends on the demand. So I want to make sure like normally if it’s on a let’s say medium to high demand product, I can get that done in one to two weeks. But if I’m having like 7,000 impressions and not that many clicks, then it’s probably better to run it for a while like we did a test one a while back of handkerchiefs. Not too many people buy handkerchiefs. This is just it. They just don’t do it. It took me three months to get data to do that because there’s just not a demand for that product. I mean, the most you’ll ever make on that products like a thousand bucks a month, 2,000 bucks a month, it was just something to play with.
What’s your criteria in terms of negative matching a keyword? Negative matching? Like how much data did you get before? Like I’m just gonna.. It’s a bleeder, I’m going to kill that keyword. If it’s eating up a lot of the campaign budgets, then I’m killing that keyword. But I’m doing negative exact. So even if I’m killing that keyword, normally there are more long tails than short tails. I mean, unless it’s completely off. So that’s how we basically choose that. I had some keyword to eat up 50% of a budget that were just terrible, right? And I’ve had other ones on the long tail they convert, but on the short they don’t. So it depends from case to case, if that makes sense. Each product is different. You guys have to understand this.
You get quite often on that when you run the report. You got other AdSense coming up? Yes.
And sometimes they’re like taking up a lot of the budget? Yes.
And I was just referring back to what you said about the AdSense not working when you put them in a positive campaign? Yes. Yeah.
If you put them in the negative campaign? It does. It does work. Okay. Cause that’s what we’ve been doing. I just wanted to double-check. But if you do that low bid on the auto, you’ll get a crap ton of AdSense and they convert really well. So you can also look at maybe putting them into like if you have AMS into the buy box. But you only have like when you’re doing AMS and you’re going after your competitor’s buy box, you use one campaign-one AdSense so that way you can really measure how that one is performing. Okay, and I’m guessing them AdSense appearing because people were clicking through from their listings? Is that right? Is that why they come up? They were clicking through from their listings or you were on the bottom of their page, right? Yeah. Other related products or sometimes they’ll actually put you under in the buy box. But that’s Amazon’s things, right? So it depends. Okay, thanks.
Yes. <question; audio low> Yes, yeah. You mentioned you bid against yourself in different campaigns. Correct. What’s the benefit of it to your viewer? Because you’re taking out more real estate, right? And you’re only paying for the click. Yeah, but you’re increasing your click because you’re bidding against yourself. So you’re competing against your one keyword in all the campaigns. So Amazon see you as (okay) one campaign, and as a campaign and you increase in your click. Not necessarily. Because the cost between the two is maybe five or 10 cents. Maybe, maybe five or two. But… So but when we’re doing that, it’s also increasing your conversion rates because they’re clicking on one of those ads so we get you more chances to get a click to get your conversions up for that keyword. So basically, it works. Yes. It works for you, you see it benefits. I see benefit, benefit for ranking. Most people want to get organic and they want to stay at the top, right? So we have a lot of bestseller badges blah blah blah and those people want to stay at the top. I might be bidding against myself just to fight off the competition so they can’t put their ads there. Okay. Does that make sense? Yeah, yeah, maybe make sense. Need to try..
Yeah. So when you take auto keyword and put it into manual campaign, will you also add it as a negative keyword on the auto campaign so you’re kind of isolating where the keyword is coming from in the auction and you have more granular control over it? I won’t do it if it’s converting well. No, because it’ll give me two different positions that way so that way I have double chance to convert and normally the manual campaign has a higher Bit than the auto.
Thank you. Thank you for the presentation. Quick question. Obviously, there’s a lot of wisdom where you know you should run PPC because increase organic and therefore your sales go up. I’m curious, have you actually measured based on because you obviously see a lot of data look through your reports the keywords that had conversions in PPC? Did you cross measure that in rankings and what kind of improvement did you see for that particular keyword? I could tell you recently we took a guy. He was, he ran out of inventory which I told no one, run out of inventory because you get punished. He was only up for about a week but he pushed the page 8 from page 1 with some waves which would be higher budgets and the exact things you can actually see a direct correlation. He’s back on page 1 in third position now after about a month. So but with high budgets and high Bits. I mean, so it was inexpensive. We didn’t even breakeven, we actually took a loss on some of it just to get it back to there.
And so, second follow-up question—do you find that targeting obviously there’s a lot of keyword software and whatnot targeting high volume keywords but works better or do you find targeting like more a broad range of long tail keywords? Long tails are always going to have a lower ACoS if you can do them correctly but you can definitely do high-volume keywords if you have conversion rates. Because if you have the decent conversion rate, then you can definitely rank on that for the organic. If you don’t have a decent conversion rate, you’ll never rank on it for the organic anyway because there’s so much competition these days at least in the US market. So I’m more familiar with US and a little bit in the UK. I don’t mess with the rest of Europe so much. Thank you.
Do you have any tips on product display ads for AMS? I’m doing really well with headline search but product display is just for in this. I think headline search is probably the best way to go. Would you agree with that one? Yeah. So she does all of AMS and she’s Marketing Director here. She’s here with me. So product display is hit or miss, it depends on the product. I’ve seen very few products that do very well like single-digit ACoS but most of the time they’re just blowing out on the ACoS so we’re doing mainly like buy box and headline search is like the best. And for product display, some people will just do it because it helps fight off the competition, right? So if they could take it on the keyword, they already have the first position they’re just dominating on that.
Is running the automatic campaign the only way you use to discover keywords that buyers use on Amazon? No, I mean there’s a lot of different ways you can use different software tools to go pro your competitors’ listings and your listings you can use trends and you can start manual campaigns with that. It’s just not as accurate or say as the automatic would be so you’re taking more guesses at that. So if I see a competitor and they’re running the same keyword for like 9 months a year on an exact match, then I can try that in a manual campaign when I initially launch and see how it does. But it has to be very similar in price or you have to be lower with more reviews something like this. You have to be competitive.
Any more questions? Looks like no for now. So say thank you very much to Michael. [applause]