insights! #98: Machine learning: who's actually controlling whom here - customer or algorithm?

Sonja Fuhrmann

Machine learning in service of the customer experience, Joubin – doesn't that ultimately make all online shops look the same?

Joubin Rahimi

Yes and no. Mega exciting question here. First: what is machine learning in this context, especially compared with customer experience? Machine learning is something like neural network. Meaning, when I input something, it processes based on information it has and spits something out. And with customer experiences it's about: "What's intention I have and how can I make it as simple or experience-rich as possible on other side? Even if neural networks all same, results will differ, cause data available in such neural network naturally differs shop to shop, and cause surrounding parameters, meaning: when do I go to shop, what am I searching, how fast am I searching, which topics do I have on that, also differ. Meaning I then have different experiences. So on one hand it's different, but all shops move toward higher level, and so, viewed from certain distance, it's quite similar again.

Sonja Fuhrmann

Okay. I personally don't know how you feel about it. I do like the personal touch in customer service. Doesn't that get lost as a result, Joubin?

Joubin Rahimi

So A: I find that much, much cooler too. Or I'm in demand mode and just get through it, but first I'd say that didn't exist yet in the digital space. There I'd like to break a lance for tools like brytes. We put information in below, because such a tool recognises digital body language. And we haven't had that so far. For example, you have a search and you type something in and get search results, and type something in again, more search results, then it'll look as if you're not finding what you're looking for. And if you recognise that, you can say: search please, filter and so on. But if you enter something like a product name, very, very detailed, or an item number, then you know you want this exact item. And then of course you can support the customer quite differently on the website. And that doesn't exist yet. And I think that then feels more personalised and more personal - though of course it isn't.

Sonja Fuhrmann

But of course there's a price for that. Doesn't machine learning lead to a surveillance culture where every click gets analysed?

Joubin Rahimi

Well first of all: the good thing about it: we're already all being monitored. We can also switch it off via cookie consent. So we already have that. But I believe more and more people will leave it on, because many things are simply so much more convenient. Regarding price, that's of course a very exciting question, because prices are also increasingly being allocated dynamically.

Sonja Fuhrmann

Really unfair. Suppose the two of us were to buy this bunch of flowers now – you might know this from home too – we'd probably both pay a different price. Isn't that completely unfair?

Joubin Rahimi

The question is, who you ask. Sure. If you ask the retailer, they say: I'm maximising sell-through or revenue or margin through this. And the answer isn't that simple within that construct. If you take, say, a Strauss, that's from a Swedish furniture manufacturer, it costs the same everywhere in Germany, unless you have the Family Card, I believe. And that's fine. But if you have prices like flights, then we all know that. And sometimes it's been done very, very unfavourably, because they say, anyone who visits with an Apple product gets a higher price than non-Apple users, people from Munich pay more than if I log in from Mecklenburg-Vorpommern with an older Windows device. That's something where profit is essentially being skimmed off. You can view that morally however you like. In B2B that's actually even quite common. On the other hand there's also quite a lot of room when I say, I'm not just looking at a single item, but I'm buying a bundle. And then it's more about people buying more, so whatever, clothing, trousers, shoes.

Joubin Rahimi

Then it's about volume.

Sonja Fuhrmann

About the volume. And then I'm also willing to simply say: okay, I'll give a better price. Then the person benefits from it, and the retailer benefits at the same time.

Joubin Rahimi

Sure, businesses are naturally after their profit. Are you not concerned that machine learning puts revenue optimisation front and centre, while customer loyalty perhaps takes a back seat?

Sonja Fuhrmann

Yes, I think revenue optimisation, we have that everywhere, and I think that's something every e-commerce lead here, or managing director of a company where the managing director works for another owner, private equity or shareholders, has to do one way or another, and already does.

Joubin Rahimi

And that's a tool where it becomes even easier. Now your question was, but does it come at the expense of the customer experience? That's important too. Yes, and I'd say, with price it's like this, if I feel it's worth it, I'm happy to pay it. If I then find out afterwards that it's somehow become cheaper, that's annoying. And there are examples right now where a lot is changing. Take Tesla, which is starting an incredible price war, especially in the car market. Some are actually parting ways with it. Sixt just said: "We're not taking any more Teslas, because that doesn't work for me." Large fleet owners, like companies, can say exactly the same thing, because they say: "I buy, and the residual value is now significantly lower. So it costs me too much." That leads directly to the B2B customer experience, but also, if you feel taken for a ride, that costs points. That's right. It's a balancing act. It's a balancing act, yes. But Machine Learning was also the question: can you use that to create experiences you didn't have before?

Joubin Rahimi

And that, regardless of price. And that's of course the exciting part.

Sonja Fuhrmann

It's exciting, and like everything in life, it has two sides. Thank you very much, Joubin, for the conversation. Thanks to you all for your attention.

Joubin Rahimi

Thanks and, as always, I look forward to comments, remarks and ideas, let's discuss in the comment field below or feel free to send a direct message.