Where’s My Mom’ is a children’s book by Julie Donaldson (she of Gruffalo fame). It follows a young monkey who is on the quest to find, you guessed it, his Mom. The monkey is accompanied by a butterfly, who tries to help out in any way he can. You have the usual games of ‘is this your mom’ pointing at, in turn, an elephant (‘my mom isn’t big and gray!’), a bat (‘my mom doesn’t have wings’), and so on.
In the end, of course, the mom finally located and hugged, to a collective ‘awwww’ (or, if you are my kids, ‘read it again!’).
Search engines can be literal
So what on earth does a children’s book written a decade ago have to do with SLI? Well, if you think about it, the butterfly is a search engine. He uses the parameters that the monkey provides him with, and tries to get the closest match. ‘Look for “MOM”’, he is told, and he looks for a mother. ‘Look for “MOM” + “BIGGER”’ he is told, and he fetches an elephant. Not from malice, or because he is dumb. This is what the user told him to search for. The butterfly is providing results based on his programming.
There’s an extended trial and error (eight animals all in all) period. And only after eight animals does the monkey provide relevant information. ‘Butterfly, butterfly’ he says, ‘can’t you see? None of these animals look like me!’. A revelation to the butterfly, of course. Because, as he points out to the monkey, HIS children look like nothing like him. As caterpillars very rarely look like butterflies, this actually makes perfect sense).
Literal search results can be frustrating
As a matter of fact, there’s no ‘can’ about it. Literal search results ARE frustrating. I’m not supposed to know exactly how to ask the question, or enter the search terms, to get relevant results, am I?
If I am looking for trousers, the search engine should damn well know that I am looking for MEN’S trousers. Obviously. I mean, look at me.
Of course, if I thought about it for a few seconds, then I’d think why on earth would the search engine know that. I might even think that I should have mentioned it in the search query. But that isn’t the best experience for me. I just want to search, find something good, buy it, and leave.
And that’s the point of the story. Providing search engines with few details can give you bad results. Adding irrelevant information won’t help you either. And it’s a rare user who says to him or herself, ‘OK, I need to phrase my search terms in a better way’. Most of us will get annoyed that the search results don’t give us what we expect. And even if we do try to search for items in a different way, that annoyance builds up. Two or three search terms in, we’ll shrug and go to shop at a different website. Even monkeys looking for their mothers only get to eight before losing their cool.
Part of the problem is Google (to take a search engine at random). Google has turned relevant search results into an art. Type almost anything into their search engine, and you’ll get a relevant result. You wonder how they knew that when searching for “rhyming guitars”, you were actually searching for this song. (Google knows everything. That’s how). And that’s the level of experience that customers have come to expect. When they come to your website, and search for, well, anything, they want relevant results.
Frustrating customers is never a good idea
During the search for his mother, you can hear the monkey growing more and more anxious. He specifically states that his mother has no wings (so she can’t be a parrot). Then he gets frustrated when the butterfly suggests a bat (“I told this damn search engine NO WINGS”). When the elephant is suggested for the second time, he finally loses his cool (“I told you she’s not an elephant!”).
The book ends well of course. After all, this is a children’s book. The monkey is reunited with his mother when the butterfly finally understands they are looking for someone, who looks like the monkey. But even then, the butterfly finds the monkey’s dad first. Because even here, the search engine was told ‘find someone like me’. And promptly forgot all about the user’s history – they were looking for a mother all the time, not a father.
SLI’s Learning Search
And that brings us full circle to SLI – in this case, the Learning Search solution. Because Learning Search isn’t a butterfly. It doesn’t accept the literal meaning of your users’ questions. Instead, it uses their search history, and indeed, other users search history. This history helps the search engine actually understand what they want to find. Be it the right clothes, the right sports gear, or even the right hardware. So with even the most basic of searches – ‘Find my mom’ – SLI can make an educated guess. In this case, that the monkey is looking for his mother. His mother looks like him. So, he is looking for (surprise!) a mother monkey.
And when the ecommerce website I’m on has a search that knows what I want as well as what I need to find – I’ll be coming back. Probably to buy the next book in the series.
