Episode 7 – Brand Analytics, Search Volume Estimates, and How We Create Helium 10 Tools

Learn more about how Helium 10 tools work, find out how your feedback can help us to continually improve our tools, and get an insider’s preview of exciting 2019 updates in our first Tech Talk with Bojan, Helium 10’s Chief Technology Officer!

In this new style of the podcast, we discuss more of the technical aspects surrounding Helium 10 tools with a sneak peek into how it all comes together from tool development to release to our members. If you have ever wondered about the rhyme or reason behind why Helium 10 has created the tools and features available to members, this is the podcast for you!

Bojan also addresses concerns that users have voiced concerning our analytic metrics, including the disappearance of exact search volume keyword data from Amazon. He also shares how Helium 10 is still laser-focused on delivering highly accurate estimations for keywords, and why search volume alone is not the best metric to use during keyword research.

In episode 7 of the Serious Sellers Podcast, Bradley and Bojan discuss the inner workings of Helium 10 tools in 2019, including:

• An introduction to Bojan and His Technology Expertise
• Bojan’s Strategic Contributions to New Helium 10 Tools
• New Brand Analytics Made Available for Helium 10 Users
• How Many Keywords is Helium 10 Monitoring?
• How Accurate is Helium 10’s Keyword Search Volume Estimations?
• Issues with Amazon’s Exact Search Volume Data for Keywords
• How Accurate Was Amazon’s Data Anyway?
• Marketing Goals VS User Goals When Developing Helium 10 Tools
• The Problem with Reporting Current Keyword Ranking Positions for Specific ASINs
• What Tools Would Helium 10 NOT Consider Making?
• How Your Feedback Leads To Enhanced Feature and Tool Development
• How to Submit Suggestions or Requests to Helium 10 for New Features or Tools
• Tool Updates Coming for Q1 and Q2 in 2019

Enjoy this episode? Be sure to check out our previous episodes for even more content to propel you to Amazon FBA Seller success! And don’t forget to “Like” our Facebook page and subscribe to the podcast on iTunes, Google Play or wherever you listen to our podcast.

Want to absolutely start crushing it on Amazon? Here are few carefully curated resources to get you started:

• Freedom Ticket: Taught by Amazon thought leader Kevin King, get A-Z Amazon strategies and techniques for establishing and solidifying your business.
• Ultimate Resource Guide: Discover the best tools and services to help you dominate on Amazon.
• Helium 10: 20+ software tools to boost your entire sales pipeline from product research to customer communication and Amazon refund automation. Make running a successful Amazon business easier with better data and insights. See what our customers have to say.
• Helium 10 Chrome Extension: Verify your Amazon product idea and validate how lucrative it can be with over a dozen data metrics and profitability estimation. 
• SellerTradmarks.com: Trademarks are vital for protecting your Amazon brand from hijackers, and sellertrademarks.com provides a streamlined process for helping you get one.

Transcript

Bradley Sutton:Brand analytics, search volume estimates, the kind of tools that Helium 10 would not consider adding. We talk about this and more with our Tech Talk with Bojan, the CTO of Helium 10.

Bradley Sutton: How’s it going, Everybody? This is another episode of the Serious Sellers Podcast. I am your host, Bradley Sutton and I am joined today by Bojan on our new series that we’re going to have here once a month or so, called Tech Talk with Bojan where we talk about different technical aspects that have to do with Amazon or Helium 10. Bojan is our CTO here, but apart from that, a lot of people don’t know who you are Bojan. So tell us a little bit about yourself. Who is Bojan?

Bojan Gajic: Hey Bradley. Hello everybody. Some of you might have had the chance to see me in the past. I don’t show my face a lot. But just in short time, I’m a mathematician by education. Then aspirations, somehow they pushed me into this programmer/managerial/data mover-around guy. So I’m here with Helium 10. I’ve been here for more than a year and I’m trying to provide value. So I hope I’m doing pretty good job with that.

Bradley Sutton: Yes, you are. So guys, if you really appreciate a new tool that we have had in the last year, like compare ASINs on Cerebro or Review Downloader. You know the new features that we have. Pretty much you owe Bojan a debt of gratitude because a lot of this, if not all of it, actually comes from him. So, he’s the genius behind the scenes here. But now he’s going to be a little bit in front of the scenes here. So let’s get to some questions that have been submitted. And one thing that we’re getting a lot on Facebook and also just support, Bojan, is about the new brand analytics data point that has been made available to certain sellers. And so people are asking us, have we integrated this into our system? Are we going to be integrating into our system? Or what can you tell us about this brand analytics?

Bojan Gajic: So it’s not–the brand analytics data is not really new. What’s new is that it was made available to a wider audience. So in the past, you had to have I believe a vendor’s account. Now if you have a brand registered account, you get access. It’s being slowly the hold out. The fact is that the data or version of that data, or some aspects of that data was already incorporated. Some of that–our tools especially keyword related tools, obviously. There are few data points with brand analytics data. Some are more flashy and and some are just more useful. We try to stay away from those flashy, shiny objects and we’ve tried to focus on useful components of it. Then, providing some value on top of what’s available there. One thing that’s sort of significant with this brand analytics keyword is that the list is relatively small. That they–Amazon gives you about 1 million keywords and it essentially turns rank of keywords by search frequency. So you have similar metrics to BSR. The potential problem with that is that the 1 million keywords is order of magnitude less than–well actually orders of magnitude less than what we have in our database. And once we manage and we try to expose to customers. So while with brand analytics data, you get the list that you can essentially import into your excel spreadsheet–you cannot do that with Helium 10 data. So they are using that data, but that alone would not be sufficient to provide the quality of service that we are providing.

Bradley Sutton: So just to give an idea, you mentioned that this brand analytics has a million keywords. I think it’s actually, it is like maximum 1 million, right? So what about in Helium 10’s database approximately how many keywords are we talking? Oh, we have 5 million and that has 1 million or 10x, or what are we talking about?

Bojan Gajic: So depends on the way you look at it, you slice it. But for US marketplace, at any point of time, we are monitoring and processing more than a hundred million keywords.

Bradley Sutton: More than a 100 million? So more than a hundred times the number of keywords that brand analytics has, we have in Helium 10?

Bojan Gajic: Right. So, we’re trying to expand that reach. So yes, you can focus on one keyword. You can focus on TV or Bluetooth headphones. So, those are the top 10 keywords and sure if you can make it on Bluetooth headphones, you will be pretty successful. But then if you can focus on long tail keywords, if you can get ranked on page one on the keyword, that’s a 2 million position on that at list, you will be pretty successful seller as well. So, we’re trying to look beyond the data that’s available to everybody, and then build on top of that. So yes, we have more than a hundred million keywords, and we’re constantly searching for new ways and novel ways to acquire keywords, and estimate the value of the keywords, and find new ways to take advantage of them.

Bradley Sutton: Okay. A lot of this has to do with mentioned like search volume in the past. That’s been a hot topic and it’s kind of leads to another question I’ve been getting recently is, you know, now that we’ve rolled out our new search volume column there in Magnet, Cerebro, and other tools. But people are asking, “Well how can I trust this search volume? Or what is it based on or why is this accurate?” And I know you’ve actually said that in some situations, the search volume estimates that we see now might even be better than some of the information that was coming from Amazon last year. Right?

Bojan Gajic: Right. So, we have that conversation here several times over the past 12 months. Some of the data points that we’re coming out of that source, mysterious source that were in the way two months ago. So some of those data points just make no sense. Flashlights, at least for awhile, or Mother’s Day’s gift. You would see search volume–exact search volume was zero for “Mother’s Day’s gift” like back in April or May of 2018. So we look at the data and it’s clearly flawed, but then again, if we tried to correct, we’re walking away from that Amazon data attributes. I wanted to–as a team we wanted to make adjustments, but our hands were essentially tied. So know that the source–

Bradley Sutton: It was garbage in, garbage out. I mean, if it was what Amazon was showing us, we weren’t going to go and change it. So if it showed us zero search volume, we also displayed zero, right?

Bojan Gajic: Exactly. So, that was supposed to increase veracity since that’s the day that it’s coming from Amazon and it was easy way out. Or they will be observed that whatever, observe, our customers observe, and ask about, “Hey, that’s what Amazon is showing.” So that’s what you get to live it. Now that, it’s ground truth was removed and now we are able to actually build our models and deploy our models, and show our estimations that are clearly more accurate than those zeros that we were showing. It’s actually a significant number of keywords. There are hundreds of thousands of keywords that returns zero exactly.

Bradley Sutton: Returned something broad, right? So that’s how we knew–definitely had to been an anomaly. Like he wouldn’t have something that’s zero exact, which is what it was returning, but then all of a sudden it would say 25,000 broad.

Bojan Gajic: And in some cases, right. In some cases, yeah, you would get broad. In some cases you had instances of what we call double zero where you get both exact and broad zero, but it’s like, it’s obvious that it’s actionable keyword that there’s value in that keyword. And just by going off of Amazon numbers, you would miss that keyword, especially that it was popular to filter by search volume. So, you would run a Cerebro search or Magnet search, and then sort by search volume or even exclude the zero, and you completely miss that keyword, or the niche, and those are good keywords. So, we were thinking about–that somehow, marking those keywords as estimates in the past, but then we were mixing estimates and actual data. So, you have to manage that public awareness campaign. So again, the easy way out was to just show what Amazon shows and stick to that. Now we have that freedom to improve on the source data. And, I believe in many aspects it’s more valuable–what we have and what was displayed when only Amazon data was used.

Bradley Sutton: Okay, excellent. So guys, in a nutshell, yes you can trust this data. These are estimates but in many cases, even hundreds of thousands of cases actually. As Bojan said, it’s actually more accurate than what Amazon was giving at certain times of the year last year. And you guys are constantly, you know, we’re not constantly maybe, but you are always looking for ways to improve these estimations and the numbers now might even see improvement going forward. Right?

Bojan Gajic: Yeah. And even when you look at that as brand analytics data, there’s a lot of noise, and there are a lot of inconsistencies in that data. So it might be a result of attempts of abuse. So somebody might be trying to manipulate Amazon data, and that’s kind of sneaking into data that’s coming out through brand analytics. But you have instances of keywords that are showing the zero conversion rate or zero sales. So the raw data that’s coming to us was–I feel it’s not properly vetted. It was just exposed because there was a lot of demand for that kind of data. So it’s still up to the consumer, or the tool builders like VR to process that data and to add value and filter out noise or junk or garbage as you said.

Bradley Sutton: Yeah, I remember the first time I saw one of those reports, you know, I was still dealing with my old company, a diet pill company, and we had access to that report. I think it was something that cost like $30,000 or $40,000 back in the day if I’m not mistaken. And I compared it to information I knew, you know, because I worked for the company. I had their sales and it was wrong. Like, now I looked historically back at the new data that’s available to everybody at those same searches that were wrong before. And it looks like they did fix itself. Kind of like similar to a lot of those search volume, not SMS, but the actual search volume that was coming from Amazon that was showing zero. A lot of those, they did fix themselves. But I think that’s something that made many Amazon sellers make the mistake of is like, “Oh, this is coming from Amazon. So, it’s 100% correct, but even Amazon, you know, can make mistakes and they fixed it. So just just keep that in mind guys.

Bojan Gajic: In many cases that it’s not really that the sellers are looking at the Amazon data directly and listening to Amazon’s marketing message. So this brand analytics data, it was not really announced, it was just sold out silently. But what sellers are seeing and hearing is that kind of marketing layer, that middleman, between Amazon and the seller. And in some parts of that middle layer, that layer between Amazon and sellers, there’s a lot of marketing noise. So some companies, and in every business is like that. Some companies might be engineering-driven, some companies might be marketing-driven, and that’s fine. In our case, we’re user and engineering-driven more than marketing driven. So some things–and that happens a lot. Like there’ll be initiatives that are abandoned or never pursued because they made sense from marketing perspective, but they did not really make much sense from adding value to users or from the engineering perspective. And they did come back to bite us a few times where we would see an opportunity from marketing angle and decided not to pursue that because it didn’t seem like good thing to do, like proper thing to do. But then you have a competing market, a competing tool, or just change in markets where somebody else would release the same initiative that we abandoned or decided not to pursue. And then they were put in a kind of a tough spot, where we have to justify why you’re not doing something. And, it’s sometimes put as our kind of technological insufficiency. So I have to defend myself fairly often. Why don’t we do this? Why it was so the zeitgeist right now, one thing that’s trending is keyword rank. So hey, can you–and we get that question a lot, like we get that question often: “Hey, can you not define even my rank, my position for a specific keyword changes?” And that was discussed months ago. And the problem with that is that position is not stable. It’s not unstable because the position changes, it might be unstable because Amazon might be testing different models and sure you do want to know if your position changes when Amazon changes model, but you might not really want to know that your position is changing temporarily for 5% of user based, especially if it can not guarantee that is 5% of user based.

Bradley Sutton: So it’s changing every single hour, 24 hours a day, you are going to get 24 emails saying it’s going up and down.

Bojan Gajic: Exactly. So, we build models and then we build kind of an internal test tools where I see emails going out. So if you’re in Cerebro, and you have a thousand keywords, And you have 10% boost. So they allow 10% of keywords to be boosted. So that’s kind of the keywords– you might have 20, 30 keywords that are going back and forth between page one and page five. So if it’s some users in some scenarios that we had enabled on the backend, I sold literally hundreds of emails that would have been sent if a feature like that is really–so we’re trying to find the balance between that marketing message but also actually making useful tools, not just something that sounds cool.

Bradley Sutton: That brings me to the next question I actually wanted to ask. So this is a perfect segway here. You’re talking about how people have suggested this or ask this, you know, as a feature. But this is an example of something that we probably would not do if we really see a clear lack of usefulness. But what are some other scenarios where people could ask for a certain tool or maybe they have even asked for tool, but at the same time we would not consider making it due to x reason. Like do you have any other examples or situations?

Bojan Gajic: Right, so they say with age comes wisdom. I’m old but I’m not wise, but I’m stubborn. So I say in my case with age comes stubborness. So, what usually happens, enough customers will ask for something that I will say, Okay, sure. Yeah. But like in general, that just tells you that’s your customer focus company. So, it’s not what sells, it’s what we think provides value. But then in some cases, obviously customers, they know much better than what works for them than what’s been up. But one thing that that’s kind of off limits, and you can ask as many times as you want, it’s unlikely that it will help is playing they’ve functionality that’s Amazon’s terms of service. And especially, I guess there was a spectrum there as well. But if things are yearly opening for abuse if you’d see very easy to make a case that the feature can be abused, then you’re not going there. So, the view matching for example that I heard was, on their consideration at some point that was never released, and it will never be used in the least. So, manipulating reviews, exposing customers’ identifiable information, or manipulating the ranking that things that are clearly against Amazon terms of service, and things that can jeopardize our seller’s position in Amazon world. That’s not something that will be considered the least. So that’s off limits for sure. Now on the flip side, we will make changes that they’ll cause some discomfort with customers. So we had– recently, we add the updates with Magnet, then one of the data points from the report was removed. So they’re returning 20,000 keywords, or 30,000 keywords. They could sort them. They can filter them. One thing that you cannot do or that you couldn’t do for a week or two with Magnet, you could not easily get search volume and for your seed keyword. So you want to see search volume for glue gun, you answer glue gun. You get 20,000 keywords and then you need to find your seed keyword in that report. So we used to have search volume for the seed keyword exposed. It was removed because our models show that it’s not being used or consumed. So we tried to provide more value if it turned out that we did not interview the user base properly so we did a survey, the terrain and we had to bring it back and we did bring it back.

Bradley Sutton: So that’s fun. That was a specific result. The reason why this came back was specifically resulted from people sending the emails and suggestions saying, Hey, we want to be able to see that seed keyword data. And so you guys will like, wow, 50 people, 60, or whatever it was are asking for this. This is something that’s, you know, pretty reasonable that we could do. And so you guys implement it.

Bojan Gajic: So, in some cases, with the list of tool with specific purpose and they’re being used in a different way. So Magnet is being used to extract search volume for a specific keyword. Does sounds really the used case that we had for Magnet. So Magnet is supposed to return a lot of keywords deleted with your seed keywords. Then you get back bunch of ideas how to expand on that keywords, or how to find adjacent niche for example. So I did not anticipate that people are using Magnet in that way and there’s significant engineering costs on the infrastructure cost on the backend if you enter a keyword just to get search volume, and we run through the process of, or retrieving 20,000 keywords in and aggregating what needs to be aggregated. They are not doing that efficiently. So, we put it back at significant infrastructure expense on our end. We will also, obviously it’s not the ideal way to do things if you have 20 keywords, you would have to run the 20 searches and somehow record those. So obviously if you’re not doing the best job–of out at meeting customer needs. So there will be some changes there, and additions that will make that kind of task easier. Moving forward, yes, we brought that back and we implemented that only for–due to customer demand. And, that’s not the only reason. So, that does not the only feature that’s being implemented because we have requests coming in from the user or users. So profits, storage fees, or even that multi-ASIN in Cerebro that we’ve been using for a few months came as a result of, kind of you request or more of a observation that be made and be talked to our users. Sometimes, it’s not just the feature request that comes from users, sometimes I will go out and talk to customers and try to observe how they’re interacting with our tools and based on their interaction, they will model our tools and modify, and try to make them more user-friendly. So that interaction is important for us. If somebody has an idea, especially if they have a problem with like friction point–is our tools or if they’re doing a lot of repetitive manual tasks or steps. Sure. We want to make your life easier.

Bradley Sutton: Yeah. So guys, you know I’ve said this before on AMAs and the Facebook groups. If you guys have suggestions for tools, you know, please get them in. We do listen to it. Almost every one of our recent updates or new tools are all coming from your suggestions because we don’t make these tools for us. You know, Bojan is not an Amazon seller. I am not an Amazon seller. We’re not making this for our own use. We’re making this for you guys, cause remember what Helium 10 is about; we help sellers rise to the top and the only way we can help you do that if we know your pain point. So again, support@helium10.com send an email, put attention, suggestions. It is going to get across Bojan’s desk, he does read these things. We just give you an example right now of an update we made three weeks ago. And then there was, I don’t want to say outcry, but a lot of suggestions about what they wanted to see differently about it. And within a week and a half of these first suggestions. Now it’s fully live on the website. So we do listen to it. Just remember there are things that there’s no point to even ask because we wouldn’t do it. And that is if it’s something that’s in clear violation of Amazon terms of service, you know, something about reviews or something that could get us in trouble. Something that could get your Seller Central account in trouble. Well obviously, we’re not going to implement a tool like that. So guys, one last thing real quick. You can’t give us any clues, or you know, I don’t want you to give anything away but we definitely have things–your team is working around the clock on some cool new updates. Some things, I mean, I think more than almost any tool that I’ve seen, we are constantly making updates. We’re constantly launching new tools. So yes or no, we do have things in store first quarter, second quarter of 2019 that people can look forward to?

Bojan Gajic: Yes. Well, we always have things in the store. So, there are a lot of, well a lot of projects in the pipeline then and just stayed with us. So next week, next month, next quarter, the one thing that’s guaranteed is this, we will have updates as well. New things and the existing tools will be improved. So sure, that’s the guarantee.

Bradley Sutton: Well Bojan, thank you very much for your time. Appreciate you coming on here. This was the first episode you guys didn’t have the opportunity to ask questions, but this will be or should be a monthly episode that we’re going to have where we get talked to Bojan. So if you guys have some questions, this is not like a technical support question. You know we have 24-hour around the clock support for that. But if you want it to ask, just a general question about how Helium 10 works or something else, go ahead and submit that to also to support@helium10.com, and then put attention SSP. That’s a Serious Sellers Podcast and then they’ll get that to me. And then perhaps, maybe it’s a question if you guys have the questions specifically for Bojan about the technical aspects of Helium 10, go ahead and submit that and they’ll get that question to me and then maybe we can read that question on the air. And if yours gets to pick, we’ll send you a Helium 10 t-shirt. Maybe signed personally by Bojan himself. He’s going to be worth a lot of money next month.

Bojan Gajic: So we were in mid February. So we’ll be in Vegas in a month.

Bradley Sutton: That’s right. We’ll be in the biggest prosper show.

Bojan Gajic: So we can do fist bumps and high fives.

Bradley Sutton: There we go. All right guys. There you have it. Thank you for listening to this episode and we’ll be back with Bojan next month for sure. Thanks a lot, Bojan.

Have more technical questions about Helium 10 tools you would like to ask Bojan? Be sure to let us know in the comments!