Why search systems fail without semantics

>> Virgil Carroll: Hey, everybody! Welcome back to the podcast. Joining, me today is Cole again. Hi, Cole.

>> Cole Schlotthauer: Hi, everyone. I'm Cole if you haven't seen me before.

>> Cole Schlotthauer: Just kidding I'm on every episode

>> Virgil Carroll: So you wanted to talk a little bit about search, didn't you?

>> Cole Schlotthauer: yeah, I think it's a pretty fitting topic. It's in the topics we cover.

>> Virgil Carroll: That's not a topic I really know a ton of things about, so I've never really done anything with search.

>> Cole Schlotthauer: Yeah, you only speak about it every, every other month or so around the world.

>> Virgil Carroll: That's true. So, but anyway, so what were you looking to talk about today?

>> Cole Schlotthauer: So, yeah, I'm just gonna get. I wanted to get into today, you know, talking about how there's so many, search systems out there that, you know, kind of suck. And that's. That's the reality of the situation here. And I'm kind of wanting to pick your brain here today on what are some of the main things that we see search systems, not doing too well at. And is there anything off the top of your head just that you want to rant about?

>> Virgil Carroll: Oh, my God, off the top of my head, I have about 500 things. but even going with some of the work you and I are doing right now, I definitely preach about a lot of different things, but one of the biggest things probably is about semantics in that and the semantics of search. And a lot of people don't really understand it in that, you know, semantics includes everything from, you know, synonyms and acronyms to how words relate to one another to how spelling happens and all those different pieces like that. I mean, you know, I always think back to when I was, quite, a bit younger, and there was like, I'm an avid reader, and there was a site out there that doesn't exist anymore that used to be like, you could put in like, one of your favorite authors. Like, one of my favorite authors is Vincent Flynn in that. And I've read all of it, all of his works. And, I can plug in Vince Flynn, and it would show me all the authors and books and that kind of stuff that were connected into it and, you know, called it a semantical web, which is basically that. Well, it's the same thing with, words, phrases, anything like that that people might search on. There's the actual words and then there's the semantics around that. All the different things that relate to it and all the different things it could mean.

>> Cole Schlotthauer: Yeah. So does having... well, first off, what's it kind of like to have a semantic search versus not having one? Like.

>> Virgil Carroll: Well, it's. It's, Pretty good. let me give you an example. I mean, you know, let's say that I'm searching on the word accounting. Okay. Just per se. And so let's think of some of the semantics around that. Accountant, finance, taxes, there's all things that are related to it. But in a non semantical search, if I just search on accounting and I had a 10 page website and there was not a single page out there that had the word accounting on it, what would it find?

>> Cole Schlotthauer: You know, probably not much about accounting.

>> Virgil Carroll: Right, Exactly. But if I was talking about like a tax professional, which could also be an accountant or somebody that helps with the books or something else like that, and I had something there and semantically, the system knew that those two things were related, then you would get better results. So basically what happens with a non semantical search is it's kind of a one to one relationship. What you're searching on is what you're going to get back. No kind of variability around that.

>> Cole Schlotthauer: And that's not how people search. So like, you know, they're not just going to guess the exact topic that you have.

>> Virgil Carroll: Right? Exactly. Nobody searches like that. And I mean I have a million examples of where that completely fails on so many levels that I've seen.

>> Cole Schlotthauer: So you're kind of just giving people like a path to, you know, search more intuitively according to like the topic that may be in their mind, but not it's not in everyone's mind.

>> Virgil Carroll: Yeah, I mean, you know, one of my favorite examples of this is we worked with a banking system a number of years ago and like one of their banks was Arizona Bank and Trust And when you actually look through and this was actually internal to the organization, this wasn't even the way people in the public see things. But you look at it, you had people that would look for ABT AB & T, AB and T, Arizona Bank and Trust. Arizona Bank, Arizona Bank Trust. I mean there were a million different variations of that and that include, you know, kind of abbreviations that people used and different phrasing that they used and all that kind of stuff. And that's just a little microcosm of it. So you know, for one name of one public institute- er one company in that you could have 50 different ways you could search for it. And that's a problem because most search systems don't take that into account and don't understand all those different relationships out there.

>> Cole Schlotthauer: Yeah, well I'm sure that's kind of a rabbit hole to get into. And we can talk about that for, you know,

>> Virgil Carroll: Days

>> Cole Schlotthauer: Many, many minutes,

>> Virgil Carroll: Years, you know

>> Virgil Carroll: I mean there's so many things talk about. But but it's, it's, it's a problem, you know, frankly, it's a problem that's out there. And it's one of the reasons that a lot of search systems really suck is because people don't really think of things. I mean, another great example is I just was Talking to my 7 year old and you know, he was talking about something sticking already sticking to something. And he said sticked. And I said no buddy, that's the word stuck. And he says well, why is that? And he said, he says why is it called stick? Isn't a stick, you know, like you find on the ground, you find a stick. And I said, well that's another thing that happens a lot in, especially in the English language. It happens in all languages, but in the English language in particular, is we use a lot of words differently like stick. Well, a stick is from a tree and it's a piece of wood. And then if you stick something, you're getting something to adhese to something else in that. And they're the exact same word with two completely different meanings. And so really what that brings it to is it's not only about all the different words, but it's also the context in which people say things. Yeah, really critical because they can say a word but have one of potentially multiple meanings for that word. And that's where when you start talking about going down that rabbit hole, now you're going into crazy land because you have to start thinking of all these. And that's where these semantics really come into play.

>> Cole Schlotthauer: You know, the English language is funny that way because you know, you can apply that same logic to so many different ways of phrasing things. And I think it's just important for a search system to be able to account for that because yeah, I mean.

>> Virgil Carroll: We've hosted a lot of exchange students in our house and that's usually the most frustrating thing is because they hear a word and it's like, well, I thought it meant this, but now I'm hearing this. And I said, well, actually it means both. Just depends on how you're using it.

>> Cole Schlotthauer: Right. And you know, one other thing I kind of wanted to talk about here was, this is kind of shifting more into the good practices of things. But you know, sometimes you'll make a search, and maybe you know, just some type of website or university, government Website type thing, and you'll get, like, the actual topic, like its own kind of, like, profile that's brought up. Oh, yeah, yeah. So can you, if you kind of know what I'm, getting at here, can you explain a bit more on that?

>> Virgil Carroll: I do. I mean, you're talking about, you know, I mean, when you start talking about how to handle this, the reality is semantics is such a big thing. And, you know, a lot of what's been happening, is everybody's looking to artificial intelligence. AI is a solution to be able to figure that out. And it's supposed to understand semantics better. But the reality is for every, every one piece of semantic, it understands better. It does things completely ridiculous on the other side and makes things related that have nothing to do with one another. So one of the answers that you're talking about is really us, kind of doing curated results, otherwise using the search systems to actually curate and kind of push things to the top. Because the reality is, and it's completely off this topic, a lot of the issues tends to be the content that's on the pages too, and that either you're not using the keywords or it just doesn't make sense to the search engine, what it really means and that kind of stuff. Or a lot of times you have pages that you think are important that have very little content on them. And so we can use things like curated results in all different search systems are called different things. Sometimes they call it promoted results, sometimes they call it, you know, result sets. I mean, all these different things. But basically it's about saying, if we get these parameters, let's go ahead and push something to the top. Let's show something different. Let's, let's take control of it. I like to call it push technology or push search, and that we're trying to push something to them. It's not that that fits with every case, but there's a lot of cases where that can really be helpful. Yeah, but the other thing is to kind of do what you're doing right now, which is word plays, you know, looking at it, taking a word and kind of figuring out all these different relationships, what are the different ways people can do it? And a great place to look at that is if you have some ability to see inside what you're getting on your searches now, whether it's an internal search or it's an external, you know, public website search, if you can see those analytics, you can really draw a lot from there and really understand, oh, well, this is the way people are searching, I mean, and that's where. And I don't want to go into this topic today because it's. It really needs to be a podcast in and of its own -  spellings! I mean, that's usually when we can just like, open the eyes and go like, oh, my gosh, there's so much to spelling out there, but there's so many different things that you can do there to really kind of make it better. But if I was to give everybody one big piece of advice to kind of say hang your hat on right here, is that the world is kind of made- the digital world has kind of made it seem like as long as you have the right search engine, you're going to be fine. Well, that's not true. Search takes work. If it takes you a long time to build your website, it should take you even longer to get your search where it's going to get. It takes time, it takes effort, it takes a lot of thought process, a lot of testing and understanding and everything like that. And the problem is that what most organizations do, which is the number one failure, is they just basically turn their search on and then they're like, well, why does our search suck? Because they've done nothing to try and help that whatsoever. And it really does have to be an integral part of the entire kind of digital process from start to finish.

>> Cole Schlotthauer: You know, it's funny, I had like, eight different things that we could have talked about today, but we ended up talking about, like, two of them. And we're at, like, 10 minutes right now so.

>> Virgil Carroll: This is a huge one. We'll definitely have to come back to this one and, look at it again. But overall, I mean, this is a great starting point. If you get one alternate phrasing or one abbreviation in there or anything like that, you're making an impact, you're doing something. And sometimes you have to look at it as that 80, 20 rule. If you can make 80% of the searches better by doing a little bit of work, it's still better than doing nothing. And that's really what we have to get to from there.

>> Cole Schlotthauer: Right? And we'll definitely be getting into a lot more of these topics here, but today was definitely a good starting point. So, yeah, thank you everyone for joining us here today!

>> Virgil Carroll: For the first part of December, where we're getting into, we're getting into the winter and everything.

>> Cole Schlotthauer: Chilly, chilly days!

>> Virgil Carroll: A great time to talk about how people can really screw, up search and that so. Well, thanks for joining me, Cole. Really appreciate it. And thanks everybody for watching.

>> Cole Schlotthauer: Thanks, everyone. Thanks, Virgil.