We were honored to team up with our friends at Salsify for their Digital Growth Series. This 8-part webinar series that provides distinct, practical advice from a series of market leaders on how brands can seize the digital opportunity ahead, rather than lose to it. Ryan Walker, our Director of Enterprise Client Strategy, understands defining the ideal tactics and strategies to optimize profits from Amazon advertising can be challenging!
But, don’t worry! We partnered with Salsify and to share expert knowledge about optimizing for profitability on Amazon. Learn how to get the most of out of your search efforts by diagnosing the health of your Amazon business.
View the transcript from the webinar here:
Hi, everyone! Thanks for joining us for the webinar today. This is Peter Crosby. I’m going to MC the webinar over the next 45 minutes or so and, if we go to the next slide, you can see that the whole series that you’re joining is really a series of eight webinars zeroing in on important topics around how brands can achieve digital growth in 2019. On the next slide you’ll see where this is really coming from is a model that’s talking about how can brands think about the different stages that they can be working, the different pillars of a growth model as you think about your activities in 2019. Their logo’s three pillars. The pillar of making sure that your fast, driving operational efficiency, what are the ways in which you should organize your processes, your people, your technology, to be able to move quickly and adjust and take advantage of all of the new channels, sales, and growth opportunities that are in front of you. Taking the air out of the processes is very important.
We’ll also have webinars that are focused on how to be everywhere and everywhere that’s important to your business. Where do you need to show up to be able to sell to your buyers and consumers? Then, finally, how do you be first? How do you get a process going of continuous optimization and optimizing every channel to be able to take market share away from your competitors by just doing it better and excelling at the sort of that hand to hand combat required to do that? Today, we are really excited to have experts from Teikametrics here to zoom in on the issue around optimizing for profitability on Amazon. I am delighted to turn the webinar over Ryan Walker from Teikametrics. Ryan.
Awesome. Thanks, Peter. First, I just want to thank you and the team at Salsify for setting up this webinar series. We’re really excited to be a part of it. To explain quickly what Teikametrics is and what we do, we’re a retail authorization platform focused primarily on Amazon, both third-party sellers and brands that are engaging with Amazon on a first-party vendor basis. We were originally founded about six years ago by one of the first third-party sellers invited to the Amazon Marketplace back in 2003 and we use proprietary data science and machine learning models, packaged in a simple software face to help you generate the largest possible sales, the best measurable ROI that makes sense for your business.
Who am I? I am not a data scientist, unfortunately, but I do like to think of myself as kind of the interface between humans and their business objectives and the data science backed automation tools that help achieve those objectives. I lead our Enterprise Managed Services Team here in Boston to support clients like Razer, Swanson, to get the most out of their sponsored products and sponsored brands spend on amazon.
Quick agenda for today. A couple things that I want to cover. I want to first go through how to understand what you’re actually getting from your advertising campaigns on Amazon. Next, understanding how much profit your business is actually going to raise from Amazon at the end of the day in terms of the total sales basis. Then, finally, concluding with some thoughts around how automation can help you succeed in both of these preceding areas.
To set the stage for the rest of this discussion, something that we believe is critical for retailers and for sellers is understanding that the growth of advertising on Amazon is probably the largest opportunity in front of us today and it’s both for people that are currently selling on Amazon or those that are not on Amazon at all. Your competitors or on Amazon. If they’re not working with Amazon directly, then a third-party seller is likely listing their products or yours and the potential lift from Amazon advertising is probably the next best dollar spent compared to other channels. We’re pretty much in the very early days of either Google Search or Google Shopping or Facebook. There’s a lot of inventory available, auctions are not as competitive and mature as other channels, so there’s a lot of space for you to come in and take advantage of that opportunity.
To give us some context around how we want to think about performance optimization in this channel, I’m going to go through some of the more fundamental details first for anyone that’s not super familiar with Amazon. If you are already an expert, please forgive me for going pretty basic to start and then, we’ll build from there, so I hope there will be some interesting things even for experts to take away from this. Before I go into too much detail, I just want to get a quick sense of how attendees in the webinar today are currently engaging with Amazon, so I’m going to launch a poll really quickly. If you could just reply how your relationship with Amazon is structured today, whether it’s a third-party seller, a 1p vendor, a hybrid model where you’re doing both, or if you’re not engaged with Amazon at all currently, I’d appreciate it.
I just got knocked out of my presentation. There we go. Sorry about that. While you’re providing those responses, I want to quickly just review some of the key metrics that are available based on how you’re engaging with Amazon. For 1p and 3p sellers you have a couple key metrics on advertising campaigns that are really important in order to understand the value of those efforts. You’ve got ideas like spend the campaign ad group product in keyword level, you’ve got revenue or sales, you also have advertising costs of sale, which is the inverse of return on ad spend. Amazon actually, just a few weeks ago, released a new, pretty interesting metric for their sponsored brands placement, called New to Brand Orders, which shows you the number of customers that buy after a click on a sponsored brands placement, that have not bought from your brand in the last 12 months. It gives you some sense of new customers. I would like to see that metric expanded to include more placements, so we can get a full picture of new customer rates on Amazon, but that’s an interesting first step from Amazon to provide us with that type of insight.
Then, there’s some metrics that are kind of unique to 1p and 3p. On the 1p side, in the advertising interface, Amazon is generally a report on retail price sales, which is different than wholesale sales. It’s important to set your ACOS target or understand your efficiency with the actual amount that Amazon paid you for the items that are being sold. What Amazon sold the item for doesn’t matter. What actually matters is what do they pay you for the item that they sold. On the 3p side, that is a less obscure where the dollars that are reported in sales are exactly the dollars that your business is collecting before fees, but there are some additional metrics that they’ll give you that you don’t get on the 1p side. You will be able to get a very clear picture of ad sales per product, while on the 1p side it’s very messy to try to piece together what was actually purchased after a click. 1p has this concept of brand halo sales, but they don’t tell you exactly what the item was that was purchased. On the 3p side, you have other unit sales and you can determine what exactly that item was.
It’s an interesting situation where the experiences across the two platforms are very different, give you a very different picture of results, but with this in mind, I think it’s also important for us to consider how incremental are the advertising dollars in sales that we’re generating for Amazon to our overall business. It is cannibalizing other channels, are these sales that would have happened organically? I’ll answer those questions in just a few seconds, but I wanted to review the results of the poll real quick, so I’ll share that now. We’ve got about 24% of those on the webinar are 3rd part sellers, about 30 are 1p, about 25 are hybrid, and about a fifth are not currently selling on Amazon at all. That’s at least the distribution of people already engaged with Amazon is not too different from what we typically see. It is a fairly even split between each of those models.
Then, to jump back to incrementality, this is kind of a classic list study example where we’re trying to understand the difference in sales from an independent variable. An independent variable here being ad spend, so it’s, I think, a bit naïve to assume that organic sales that you’re generating on Amazon today will always be there, but for a moment, we’re going to kind of assume that that is the case. What we’ll find is in an example where in the controlled portion or period of an experiment, we may have generated… Looks like we’re still actually looking at the results. Did it flip back? Cool. Sorry about that, guys. In the control period of an experiment, we may have generated $250,000 in organic sales by adding . I think the sharing… yep.
Perfect. All right. To jump back to the experiment, I apologize about that. You’ve got a case where without advertising, you may have generated $250,000 in sales. With advertising, you may have generated now $300,000 or $320,000 in sales. A portion of that probably would have occurred no matter what and the difference in that value is your true incremental results. That’s the type of thing that we’re trying to measure, to make sure any ad investment is actually producing top-line and bottom-line growth. It’s not just taking a portion of the control bar that we would have gotten anyway. A way for us to take this a step further, not every tactic in advertising is going to yield the same uplift. If we think about different ways that we can advertise on Amazon, perhaps like branded search, not branded search, competitor search. They each yield different levels or degrees of uplift.
Branded search, you might eat into a lot of your organic sales or a lot of the sales that would have happened in the control of an experiment. Non-branded search, a little bit more incremental, not perfectly, and competitor search, that might be totally incremental. Each of those has a different degree of uplift and what makes this a little bit more challenging is like you’re not running these very clean tests one at a time. You’re usually doing all these tactics at once, so it can be a bit hard to unpack what’s truly responsible for driving growth in your advertising spend and results. The question that I want to try to answer is how do I know what I’m actually getting from my advertising with this idea of incrementality in mind?
What we like to use is this idea of TACoS. What we mean by TACoS is Total Advertising Cost of Sales. You can calculate this by dividing your total spend by your total sales. It’s very similar to the way that you would divide your ad spend by your ad revenue to calculate ACOS. This is kind of just representing the idea of portion of total sales from ad spend. In isolation, TACoS doesn’t really tell us anything about incrementality, but when we measure it over time, or relate it to changes we’re making in our advertising campaigns, it can help us understand what impact to our total revenue we’re getting from advertising. To dive a little bit deeper there, what we would like to do is look at the relationship between TACoS and changes in spend distribution across those tactics I had mentioned earlier. As I invest more heavily in brand search, what happens to my total sales or as I reduce spend in non-branded search, do I find that my total sales are decreasing dramatically?
What we don’t like to see is TACoS going up while ACOS is going down. What that would indicate is advertising is generating increasingly large proportion of your total sales and your total possible profit at the end of the day is also being reduced. What we really want to see is your TACoS either staying flat or going down as you increase the rate of advertising investment. If you see that as an indicator, that means any changes that you’re making to your ad campaigns are yielding directly incremental results and you’re almost certainly not cannibalizing any organic sales that would have occurred without advertising.
This is just one way. Probably the easiest way to understand the rate of incrementality for your advertising campaigns. There are other ways you can do it and I want to talk through a catalog split that we recently did as well. This is not a perfect lift methodology. Amazon doesn’t make it easy to split by users and do very clean hold out tests, but we find that a lot of brands are sometimes unable to devote the appropriate amount of advertising dollars to Amazon for any number of reasons. You can kind of use that as an opportunity where if you can’t advertise every single one of your products to the extent that you want to, you can perhaps set it up in a way where you intentionally withhold a certain set of products for a period of time and compare the results of that product set to another set that you did advertise.
What we’re looking into test in this graph is basically an index amount of units sold per ASIN per day over time. The gray line is the non-advertised units, the purple line is units that we started advertising back in November and the height in those two lines is really the percent difference in units sold as a result of advertising. When you do this type of analysis, you can get a very clear answer to the question what would I have sold if I didn’t advertise at all? If you have a CFO that really needs to know what this number is, and the TACoS methodology isn’t sufficient, you can split up your catalog in this fashion, but again, it’s going to have some degree of uncertainty because of just the tools and data that Amazon both gives you and allows you to use in your targeting.
From this idea of incrementality, what we also want to consider is how much should we be spending? There’s a couple of ways that we can approach this problem. You can use a variety of benchmarks from other businesses that are similar in type or size to you. You can look at your financial results and understand what kind of cost tolerance remains after all of your fees and cost in manufacturing to invest in advertising. I’ll go through a couple of those examples in a little bit more detail. One of the benchmarks you can use is just a pretty simple distribution of spend by campaign type, so we like to recommend that on the sponsored product side, you run what we call a three to one campaign structure where you have three manual campaigns for every one automatic campaign and usually those campaigns are related to just one product or one very closely related family of products.
For a little bit more detail, the automatic campaign is really designed to drive keyword discovery for you, so it’s not meant to be a key profit driver, and it’s not going to be the most incremental tactic that you’re using. We like to see this be between 10% and 35% of spend. That’s really where we see optimal efficiency in terms of most brands ACOS targets. On the branded search side, so when people search for your brand name, and you bid against that, that search term as a keyword, we don’t want to invest all of our dollars here even though it might be the most attractive place because of how efficient it is. Our feeling is that, and this is generally confirmed through hold out studies, is that this is probably your least incremental revenue dollar from search advertising. We like to see this be between 10% and 25% of spend. You should be there, you should be playing defense, and you should be getting those sales, but it’s not where I would encourage you to devote most of your time and attention. Really where you want to focus most of your time is on the non-branded side of the generic category terms that people could be searching for to find your products. That’s going to be one of your most incremental sources of sales, and it’s also where you can invest to the broadest extent possible.
Finally, competitor search should be about 5%-10% of your spend. You add those up, it’s a little bit less than a hundred, so there are a couple of other areas that we like to recommend in terms of spend distribution, so there’s also these concepts of sponsored brands campaigns and product attribute targeting that Amazon recently rolled out in the last couple of months. We generally find, in sponsored brands or headline search as it was formally known, that if you’re spending about 5% to 10% of your overall budget on this placement, you can run it at an ACOS level that’s at the equivalent or close to your campaign average. Unless you’re very interested in just buying a lot of brand impressions, which at certain points in a business lifecycle might make sense, unless that’s a goal, we would typically find 5% to 10% of spend makes a lot of sense from this placement.
On the PAT side, there’s the remaining 1% to 9% of spend. This is where you’re either targeting ASINs, or you’re targeting rule-based categories of products. Usually you can use this to target your competitor products and get a placement on a product detail page or if you just want to show up in the search results for particular categories, and you don’t want to do a whole lot of keyword research, you can just use the keyword or rather the category targeting rules. I would split that up into three different campaigns. I would have an ASIN only campaign, a category rules campaign, and then a defense campaign where you’re actually targeting your own ASINs and just trying to get on the product detail pages. That last one, I would devote a very small amount of spend to because it’s strictly defensive.
In terms of as a percent of whole, we typically find that brands and sellers are spending somewhere between 5% and 35% of their total sales on advertising. This kind of depends on the skill of your business, and it will also depend on your margin profile. For instance, if you only have 10% of your revenue left as potential profit after fees and COGS and any other expenses, it probably doesn’t make sense to spend an additional 20% of your total sales on advertising. Instead, it would probably need to be below 10% in order to generate any profit. Understanding what the tolerance is for profitability is key when setting up a budget. With that said, this is kind of just a general rule of thumb and should change depending on the lifecycle stage for products in your catalog and then, also, your overall brand. If you have a brand or a completely new product, you do want to be a lot more aggressive in the way that you invest for that product. You probably do want to be a little unprofitable. To give you a little bit more color on that, we will recommend a matrix of targets and margin ranges based on the lifecycle stage of your campaign or your product.
For products that are very new, you’re going to be a lot more aggressive, set much higher ACOS targets. For products that are your profit centers, you want to be a lot more conservative in the amount that you invest. You still want to invest some, but you’re not trying to create a flywheel effect where you’re buying ads to generate initial sales to create sales rank and reviews that generate higher organic rank naturally. With your profit, you are just interested in maximizing the profit dollars, the margin dollars, that you can capture from those items.
How does this idea of setting ACOS targets relate to overall profitability? I’m going to break down how you can think about profit for 3p and 1p really quick. On the 3p side, if we consider an example of a product that sells for a hundred dollars, there are a lot of areas where margin is taken out of the equation. In this case, let’s assume we’re spending $40 to manufacturer that item. We’ve got about $15 in Amazon fees, another $20 in FBA fees, and then we’ve got about $10 in ad spend for this particular product per unit. That leaves us with about $15 remaining gross profit. That’s pretty healthy. What you want to make sure is happening, though, is that you’re not spending more than the difference between your sale price, and your COGS price in your fees and ad spend. Always making sure that for more mature products, you’ve got some degree of profitability remaining. Amazon makes this pretty challenging. You’ve got to go to a whole bunch of different reporting systems in order to get to each of these details, but you do have at least on the ad side, a very clear picture of what the actual revenue is.
On the 1p side, price is harder to control than 3p and there are other allowances similar to FBA fees or just general selling fees. Like on 3p, like base accrual, which is just for non-advertising marketing expenses or Amazon allowances for things like damaged goods and freight. You’ve got a similar equation to generate what the gross profit is per product. What I’ve found though with brands on the 1p side is that generally the teams that are responsible for selling these items doesn’t have perfect visibility into the COGS. I would really encourage you to try to get this data if you don’t have access to it already to make sure that you’re driving real profit. If you don’t have it and you want to just do a quick gut check of your profitability, you can just look at total ad spend divided by total units sold to get an ad spend per unit. Then, subtract this from the wholesale price, use roughly 10% to 15% for total Amazon fees and if there is a lot of dollars remaining, you’re probably in good shape. If there’s like a small percentage of dollars remaining, might be time to check with finance just to make sure that you’re actually generating profit from these items.
In order to know what your gross margin is to inform a sound ACOS target, you need to aggregate all these data sources together. As I mentioned, Amazon does not make this very easy. One thing that we believe is really important is having clear visibility into this. Then, beyond that, once you get a clear idea of your profitability per product, applying intelligent automation to make sure that you’re remaining profitable is going to be critical.
The next question we’ll ask is how much profit am I making per product in real time for my entire catalog? An example we just went through is great, but more importantly, I need to know what am I actually generating, how much can I actually spend on advertising at the product level? This brings us to a point where we went in a reverse order. We started with advertising and then went to profitability. In reality, you should probably consider profitability at the product level before you start advertising. If you’re losing money from a specific mature product before you’ve even started advertising it, probably not a great candidate for any amount of advertising investment. Probably need to spend time working on the sourcing or the packaging to reduce the cost there. Then, once you’ve got your set of products that you can generate a very clear profit from, that’s when it would make sense for you to turn your attention towards advertising.
Then, to bring this back to automation, the key piece that Teikametrics has is that we believe human assisted software is key to success. What’s most important to me that is that you commit to relying on automation to take you where it makes sense. If you’re not already using some form of automation to optimize your advertising or help you understand profitability, I strongly encourage you to figure out a way to incorporate this into planning for this year. To go a little bit deeper, there are a couple of important questions that we can ask ourselves to ensure that we’re maximizing incremental profit.
For instance, a typical campaign that we manage may have nearly 200 campaigns or maybe 10,000 keywords. How are you physically managing that extent of an advertising effort? Are you going in every day, updating days at the keyword level? Are you regularly creating campaigns on your own manually and how much time are you spending doing this? What frequency are you doing it with? From that, how incremental is that time spent? You’ve got the notion of incremental ad dollars or incremental revenue, but there’s also the possibility that your time could be better spent doing other incremental activities. If you’re making slower and frequent changes, you’re also missing opportunities to create sales that could have occurred otherwise.
When it comes to automating advertising, one area that we feel is critical is for humans to set parameters for computers to optimize within, so this is kind of that interface between humans and software. Humans kind of need to tell the computers what are the rules that we should be optimizing within. A machine isn’t going to be able to tell you what your ACOS target should be or what your profitability level is. We’ve got to think a little bit critically about that. Consider things like what’s the likely incremental value based on the tactic type that I’m engaging in. Am I doing a brand campaign, a non-brand campaign, competitor campaign? What’s the product life cycle stage? What’s the overall business objective, like are there certain levels of profitability that the company’s striving for? Consider all of that and then, set targets at a tactic or campaign level with the idea of product age in mind or with the idea of how incremental is this particular campaign in mind? Tell the machine, “I want you to find a way to generate for me as many sales as possible at a 10% ACOS because that’s how much profit dollars are remaining for me to maximize my profit on this item.”
It’s really challenging to do this at the scale that Amazon allows. Again, we go back to the example where we’ve got a campaign with 200 different ad groups or 10,000 different keywords. It’s really challenging to do this for all of the different permutations of situations that you could be in. There isn’t just like one ACOS target that makes sense for your entire catalog. You’ve got to get really granular and think about the best way to approach each of those items and tactics individually. That’s really where the value of automation comes in.
One of the things that Teikametrics spends a lot of time doing is again, this type of bid automation. Understanding what the right amount is to be bidding at the keyword level, at the product level, at the tactic level, based on the advertising goals that you set. Then, it’s constantly in almost real-time updating those bids to make sure that we’re at the optimal level. This just allows you to devote your time to much more important things. You don’t need to be worrying about, oh, shoot, did I update my bids in that campaign that I launched yesterday? All that’s kind of ticking for you by the ACOS target value that you tell the machine is critical.
Another part of this is beyond bid automation. It’s important to make sure you’re spending the right amount per keyword, but also ensuring that you have the correct keywords in your campaigns further enables your success. You can have the best bid algorithm in the world, but if you’re not mining the right keyword through things like automatic campaigns or manual keyword research, then your results from that algorithm are going to be muted. There’s a couple of ways you can do that. You can either move keywords through levels of refinement, so finding search terms or keywords that are working really well as a broad term, and they’re matching a more complex search phrase, creating a new keyword with the match type of phrase could add a lot of value going down the path of and moving things from phrase to exact as you find more precise and effective queries that humans are searching for to find your products.
You’ve also got keyword negation that machines can help us with where identifying keywords or search terms where you’re not generating results could be just as valuable to generating profit as finding new keywords. Saving money in areas where you’re not going to generate a sale, it might even be more incremental and more profitable in many cases than just finding new areas to invest in advertising. Then, again, isolating keywords from automatic campaigns that are performing well, bringing them over to manual campaigns, kind of kicking that flywheel off again, was another way that you can increase the value that you’re creating from your advertising campaigns through automation.
Using machines that power this allows us to invest our time in more value-generating activities. Yes, you can spend your entire day doing keyword research, fine-tuning your bids in bulk ops based on value for click rules or you can spend your day researching the next product, designing its packaging, sourcing a new item from a new manufacturer, optimizing the content on the listing page. These are all things that humans are much better equipped to accomplish than machines and there are things like bid optimization and keyword optimization that computers can do much faster and much more effectively. This allows you to invest your headcount in areas that can create the largest long-term value for you. This is where Teikametrics and my team really strives to create a lot of value for our clients. Getting this flywheel spinning of driving the best possible advertising results, creating more sales, and improving sales rank, getting more views in a self-reinforcing virtual cycle. You don’t need to spend a lot of your, in physical labor time, optimizing that if you have an intelligent bid and keyword optimization system assisting you.
Just to review quickly. It’s important to think critically about the way you’re measuring performance and what makes sense for your business. It’s not enough to just look at the ad sales that Amazon is reporting, but understanding how incremental any additional ad dollars are to your overall business is necessary. Profitability’s important and making sure that your advertising items that can be profitable is key, but the degree of that importance is going to depend and vary over time. It’s going to depend on is the product a brand-new product or is it something that you’ve had around for years and people buy it from you regularly? How much you need to invest in those two situations is definitely going to vary. Finally, automation is critical to success. You could spend all of your time thinking about those two items, or you could spend all of your time finding areas to allow machines to worry about those two items for you as you expand your catalog and expand your efforts on Amazon.
With that, I’m going to turn it over to questions quickly. Definitely would welcome any thoughts or questions or any clarification you guys might have.
Hey, Ryan. Yeah. This is Peter. We have questions coming in, which is awesome. Keep them coming, guys. It’s in the question panel of your GoToWebinar control panel. Throw them in, we’ll get to as many as we can and those we can’t, I’m sure Ryan and his team would be willing to answer through follow-up emails if that’s… Ryan, I just made a promise for you. Is that okay?
That’s quite all right.
Okay. It would be hard to say no, now, wouldn’t it? Sorry about that, but I know you’re very helpful. Yeah. Let me start out, we have one that’s kind of a simple one. The catalog split test report, someone’s asking how do they get that.
Yeah. That is a report that you’d have to generate. That’s not something that Amazon provides you with. What you’re going to do there is, on the 1p side, actually really, really annoying, so if you’re interested in doing this, I apologize because I’m about to send you on a pretty interesting goose chase. In ARA, you can pull down the number of units sold per day. However, you can only do this one day at a time in ARA Premium or you can do it one day at a time for only the last week in ARA Basic. You basically need to do that for your entire test period, and then you can flag basically each ASIN as advertised or not advertised. Then, you’re going to index the units sold per day against an average. I can send you a template if you want to follow-up by email. I’m happy to provide that as a follow-up to this.
Great. We have a question from Brendan. “How do you isolate external factors when trying to measure or test incremental sales?”
It’s tricky. I would assume that Brendan’s referring to maybe other traffic sources that are pointing to Amazon. I think it should be true that any external factors should be impacting both advertised and not-advertised products simultaneously. Now, if you kind of like stacked tests where you’re only sending traffic to certain products on Amazon, you could color your results in a way that you don’t want. I would recommend that you just be careful about external things that you can influence outside of Amazon and how broadly that effect might be occurring. For things that are outside of your control, I would believe that it should be impacting advertising on advertised products equally.
All right. What I’m going to do, Ryan, because we have a bunch of questions about your product in particular, so I’m going to sort of try and roll them all up into a big ball of questions that maybe you can address. Then, if for some reason we can also answer them individually offline.
People are asking things like, is Teikametrics fee structure based, somebody using one and a half million on Amazon annually across 150 ASINs, what do you see the main benefit to choose Teikametrics over others? Let’s see, can your software connect to all visual advertising platforms: Amazon, Google, Facebook, eCommerce site? Does tool, does it work for 1p vendors as well? How many ACINs should we have to make flywheel necessary or rentable. I’ll stop there. I felt like they were all sort of around the capabilities, pricing, and the applicability of your product.
Sure. I’m going to start with scope first and I’ll kind of work my way around all of those questions. In terms of what we can provide strictly on Amazon, strictly for sponsored brands, sponsored products, yes, we can support both 3p and 1p. 3p’s functionalities are a little bit more robust just due to the amount of data that Amazon provides through their API’s. The 1p side, a lot of the total sales data is still gated behind ARA without any real way to pull it out systematically. Some of the reporting features that we provide on 3p are not available on the 1p side, but in terms of bid automation, keyword automation. All of that works incredibly well. In terms of the number of ACENs that I’d recommend, you could start with… I’m personally managing campaigns that have as few as seven all the way up to hundreds of thousands. It really doesn’t matter. The bid automation is going to be effective no matter how dense or sparse your campaign structure is. I would also believe that even for really thin catalogs, you’re still going to have a lot of keywords that you’re bidding on and you could probably even get more sophisticated in the way that you build the search terms that you’re trying to target, so bid automation is just as important in that case.
Then, on the fee side, there are a couple of different ways that companies can engage with Teikametrics. We offer both software on a self-service basis, if that’s something that you’re interested in and you’d like to support your teams that are doing advertising on Amazon with a really intelligent platform. We’ll also offer managed services, so if you don’t want to have to worry about managing your campaigns, creating your campaigns, doing the keyword actions, we can handle all of that for you. The models that we’ll generally do for those is typically a base fee plus a percent of spend. The software is usually quite a bit less expensive than the managed service.
Did I answer all of those points? Was there anything else related to Teikametrics that I could provide color on?
No. I believe you got most of the ones that I grouped together and I want to make sure we cover a wider array of topics as well. There’s a question from Camille around what is your take about the new bidding strategy and adjust bids by placement when you are using rule-based bid automation?
Yeah. Just to be clear, we’re not using rule-based bid automation. We’re using keyword and product level bid automation. We’re trying to predict the expected value of a keyword at the product level to generate a bid value that makes sense given an efficiency or margin objective. It’s not totally rules-based. To provide our thoughts on the new bidding types that Amazon offers, I guess I want to just describe how there’s kind of two new ones. You’ve got a replacement to bid plus, which allows you to target, you know, allows you to bid more or less per placement. You also have like the up/down bid modifiers that Amazon also rolled out recently at the campaign level where you could either allow Amazon to bid up or down depending on their predicted conversion rate for the user. You could allow Amazon to bid down only based on the predicted conversion rate or you could run fixed bids and on the placement side, you could bid up for top of search or the first page, or you could bid up on strictly product page results.
I’m a bit more interested in the placement modifiers strictly because depending on what type of tactic you’re trying to engage in, it gives you the ability to kind of isolate which placement you’re trying to win. If you’re running a campaign, like a PAT campaign, you only want to be on product detail pages, it makes sense to use the placement modifier to just be on those pages. Bid a little bit lower overall at the campaign level, keep the top of search bid modifier at zero and then crank up the placement, product detail page placement bid modifier. That’s generally how I think about it. I wouldn’t use it wholesale. We generally like to, as much as we can, try to predict the expected value at the keyword level independent of where it might be placed.
Then, the other campaign level bid modifiers we’re a little bit more skeptical of. They could be useful, like having some notion of predicted conversion rate could be really valuable, but Amazon hasn’t really exposed any information about like what the results of that are. You’re basically just handing them the keys to handle your, really some of your bids and they’re not really telling you what they’re doing with those keys. It’s difficult to know how you should respond to changes in performance when that’s going on.
Great, Ryan. I don’t want to overstay our welcome. We have tons of engagement in the platform, so first, since we have so much engagement, I wondered, Ryan, if you wanted to share your email address, just if people wanted to follow-up directly with you and then I’ll throw you the final question.
Okay, awesome. My email address is RWalker@Teikametrics.com. No dot, no space, just RWalker@Teikametrics.com. The name of the company is here in the deck. I believe these slides will be sent out afterwards as well.
Yes, they will. Yeah, we’ll be sending out the presentation, as well as the recording, so no worries there. We just heard from Jason. “This was great. Well worth the hour.” Thank you. That’s really awesome, Jason. Thank you. It’s great to have that feedback.
For the final question from Kim. “How long do you see companies spending advertising dollars, creating a loss, to help new items get traction? Can you provide more guidance for his practice?”
Yeah, so I’m not going to encourage you to spend an infinite number of dollars to do that. I would probably wait until you get your first half dozen or dozen reviews on the item and you’ve got a decent sales rank before bringing the ACOS target back down to a more standard profitability level. There’s space in between those two points, so you don’t need to go from a break-even or even net loss at the product level straight to a 50% profitability rate on the item. You can use almost a step ladder down to get to that profitability point and if you see your TACoS going up while you’re doing that, that might indicate that it’s not time to bring that ACOS down just yet. TACoS is like a good guiding metric when you’re making these types of things and you’ve got very clearly isolated variables that you’re testing. Just look at your total revenue and if the total revenue at the item level is softening at a rate that’s greater than what you’d like to see, be a little bit less aggressive in the way that you’re pairing back your advertising investment. I know that’s not like a very precise formula, but test it and see is probably the best approach.
Terrific. Everyone, thank you so much for, first of all, Teikametrics, thank you so much for being on this series and for sharing this expertise and thank you everyone for joining us, taking time out of your day. As I said, we will be sending out the presentation, the recording, and you have Ryan’s email if you have specific questions. We hope to encounter you on the rest of the Digital Growth Series. Thanks everyone. Have a great day.