Interview: Voice Search Is Not Completely Changing eCommerce Just Yet
Originally published on Omniconvert eCommerce Growth Blog
Our CEO, Antanas Bakšys was invited to participate in Growth Interviews, the fun, stimulating and engaging series of conversations driven by digital business growth.
Antanas is passionate about developing innovative solutions that create value for big eCommerce companies, as well as having the potential to change the way things are traditionally done. Since 2009 he has sold one eCommerce company (GetPower.LT) and worked on several projects as a freelancer.
Voice search and machine learning are two buzzwords that have made quite some headlines in the eCommerce world lately, since it seems everyone thinks they are changing the game. But is eCommerce really that different from what it was several years ago?
During the interview, they asked Antanas what his opinion is and whether he thinks these new technologies are reshaping the eCommerce landscape. His thoughts, along with some growth tips for eCommerce marketers, can be found in the video below.
The full transcript of the interview was prepared by the authors so you can easily come back and search for any tips you might have missed. Please, find it bellow.
Why It Is Important to Correctly Interpret Your Data
Who are you and how did you get into the digital landscape?
Antanas: ” My name is Antanas Bakšys. I’m the CEO of SearchNode. Six years ago, with two of my co-founders, we started SearchNode and it was our…I wouldn’t say hobby but it was…our side-project, because we saw in eCommerce that site-search is a huge challenge. We had an online store ourselves and people were searching for stuff there, but they didn’t find anything. And we saw that, well, it was killing their conversion rates.
So, we started six years ago, and we’ve been rolling from there. Actually, back in 2017, we realized that we were just another search API. We actually didn’t have a lot of competitive advantages, so we transformed our business. We were always improving, and right now, we are kind of reshaping the eCommerce search landscape in Europe and also moving into the US. “
What frustrates you the most about the digital landscape?
Antanas: ” There are a lot of things. There used to be the things with blockchain, which I was really frustrated about. You know those ICOs, but it’s not really related to eCommerce nowadays. Actually, one thing that really frustrates me – or that I spend a lot of time thinking about – is when everyone says that you should make data-driven decisions. Everyone says that data doesn’t lie, but data lies a lot.
There are still a lot of eCommerce people, eCommerce merchants, that are trying to make data-driven decisions, but they make the wrong decisions because they don’t interpret the data in the right way.
Therefore, the question that I’m always thinking about is what KPIs, or what metrics, should every eCommerce marketer or eCommerce business developer track, and how should they interpret them? I think a lot of education is needed here. How to understand the data, and how to make decisions out of this data – this is, right now, my main frustration.”
Top Growth Ideas for eCommerce Marketers
Can you give us some growth ideas for eCommerce marketers?
Antanas: ” I can give those ideas from my point of view, from the angle I’m working on, which is search. I’m not going to talk about how to get more traffic and better traffic. When you’ve got this traffic into the online store, what the user is doing is searching for the product – one way they are searching is navigating and just clicking somewhere, or they are putting some search queries into the search engine.
One idea would be that, if you are in Google Analytics and you are right now, for example, checking your top 20 or even top 100 search queries, this is just a head; if we take the head of a long tail, it is still just a head. If you are optimizing your search for the top 100 search queries, usually you’re just playing with about 10% of all the revenue that’s coming from search. Therefore, I would really recommend analyzing more search queries – the longer search long-tail – to analyze what exactly users are searching, where it could be improved, and simply calculate the percentage of revenue coming from the top 100 search queries. If it’s 10, 20 or 25%, while you still have a long tail, you can optimize for it. This might be one (and I would say, the most important) idea.
There are other areas for search growth that you can check, for example the very basic stuff, and many eCommerce companies don’t know how much revenue this generates users of search. They were about…I don’t know, 5%..3% percent…but we have clients, especially those who are getting a million visitors per month or more, where users will use site-search internally. They generate around 70% of all of the online revenue. Search, in their journey of purchase, is extremely important.
Once you understand that this is important, when you analyze those search queries, not only the top 100 but also those that are not very frequent, and when there are a lot of them in the long tail, then you can see where it might be improved. Usually, you can find some low-hanging fruit there, just some stuff that you can fix immediately. Sometimes, you’ve got to improve algorithms, but other times, when they’re standard, you will need external help, such as SearchNode or some other companies. “
Tell us a success story from your experience in this market.
Antanas: ” I don’t want to promote SearchNode here a lot. I’m thinking about something more universal. We started working with Decathlon in Poland, a big sports company. They had their search and they did it in-house. They had some technology, and their people were improving their search themselves: checking Google Analytics, trying to do some redirections, writing some synonyms…doing this stuff.
The main idea was that it required so many improvements that there was this manual work like putting synonyms into the search, or redirections when there is a search query, to redirect to some categories, that was not scalable. They were making 10 or 20 changes per day.
When you are selling tens of thousands of products, when you have, let’s say, hundreds of thousands of search queries per month, there should be algorithms that take care of your search, because users are putting in text differently. You just need to optimize, no matter how users search. They can search, for example, in English, knife or knives, and in Britain, it is totally different.
The success story was that we started working with them, using our solutions and our algorithms, and we optimized a lot of search queries, which previously they hadn’t thought about or didn’t have any chance to optimize. We also worked a lot with their data processing and the result – actually it was so crazy I couldn’t believe it myself – was that all search improvement increased the conversion rate of users who used search by 33%. My first reaction was, ‘It’s not the right data! Let’s take a deeper look.’ However, it was the right data.
This is a huge one. I cannot say right now, not everyone will get the same result, but looking at the search…for example, where we are right now, in Barcelona, in the eCommerce Summit, I have talked to many eCommerce retailers from smaller to biggest, like Nike.com. Most of them have a lot of potential in search.”
An Insight into Voice Search and Machine Learning
What do you think about voice search?
Antanas: ” That’s a very, very good question actually. I think about voice search from two perspectives. One perspective is when you talk to some device like Alexa. I think it’s cool; it’s trending. I do not have data myself, but it looks like it’s quite useful. There are articles that no one buys from Alexa, but I don’t know. In this case, I think this will grow and it will go in the right direction.
If we talk about the word-search within the online store, when users interact with a search box, we have clients right now today – maybe 20 websites – where we have implemented voice search such as SearchNode. First of all, the usage is super low, I can tell. The benchmark is about 0.1-0.5% of all the search queries that are made by voice. Search queries nowadays are still quite simple. Users don’t say, ‘I want to buy real nice shoes for running a marathon.’ They just put in “marathon shoes” or “running shoes”, so the behavior is the same right now, since the textuals and metrics like click-through rates and conversion rates, are actually very similar to the textual search.
So, for now, I don’t see any metrics and I don’t see any trends that suggest voice search and eCommerce within the eCommerce sites would be something big. This is my point of view. If it happens, however, online stores should have a smart search, so if the user enters, ‘I want to buy a cheap phone.’ it should understand that – but later, when it’s done, because right now they’re not using this kind of query.”
What’s your take on machine learning?
Antanas: ” When I’m in the eCommerce market, I see that people use AI and machine learning more as a marketing term rather than the technologies. We sometimes get inquiries from our potential clients who are asking, ‘Hey, where’s your AI?’ We say something like, ‘Well, you know, we use some machine-learning models in some of the areas, but it’s not the AI that you understand, it’s something superficial. It’s that our algorithms, using data, just optimize things.’
Generally speaking, machine learning and AI are definitely the future. In eCommerce, it will help to automate a lot of stuff that people are not able to do manually or by some statistical models when you need machines learning from the data and making the decisions. That’s definitely the future.
For every eCommerce retailer, however, what I would recommend is to dig deeper and see where the real value of this AI or machine learning is. They see their companies saying, ‘We are super AI’, but it’s just a statistical model, simple automation, and nothing related to artificial intelligence. Right now, they’re just marketing buzzwords. For example, in search, we use some machine learning models but we never say it’s AI-powered search, because it’s not AI as it should be. Maybe in a year or two!”
Company Transparency Is a Key Success Element
What would be your advice to yourself when you started this company?
Antanas: ” I’m not sure if it would be relevant for eCommerce merchants but for software-as-a-service companies, it’s all about branding! When we started SearchNode, we were focusing a lot on product development like trying to get the best products, the first clients. We had really been a reserved brand. We didn’t tell the market anything about what we were doing. We were not transparent at all!
Right now, after six years in the market, we’re quite well-known here in Europe, but from time-to-time you’re just going to the US or talking to some other European countries, and they’re like, ‘Hey, why haven’t we seen you anywhere before? You’re doing a great job!’ So, I would say, branding is the thing, and providing relevant content and transparent information about the company, would be my advice that we needed to follow earlier. And, I think, for eCommerce merchants that’s also important.
Even myself, when I’m trying to find something and buy it, I sometimes find some online store. It looks nice, but I’ve never heard of it. There’s no information about them, such as where they’re located. What’s their story? On LinkedIn, they’re just two people. I’m not sure if I want to give my credit card information to them!
I think it’s all about branding and being transparent about who you are. What are your questions? What are you frustrated with right now? What are your main challenges right now? This transparency would be my advice both to me and, probably, to some other founders of software-as-a-service or eCommerce companies.”
End of the interview.