Paywall business models for publishers rethought

Christian Kallenberg [00:00:02]:

Welcome, dear listeners, to the new episode of the Audience Development Deep Dive by Sprylab Technologies and We Like Mags. One of my first guests on the We Like Mags podcast was Benjamin Kolb, CEO of sprylab technologies, with whom I talked about audience development and who is sitting across from me now. At the end of our conversation, Benni and I had the feeling that not everything had been said about the topic, and that's why we decided to dedicate a series special to the topic of audience development, in which we would interview guests together about the various aspects of this increasingly important topic. And, Benni, I'm really, really grateful that you, as a digital expert for publishers, are co-hosting this format with me. And tell me: What are we talking about today?  

Benjamin Kolb [00:00:53]:

Hello Christian. Our guest today is York Walterscheid, Managing Director at CeleraOne, market leader in Germany in paywalls and monetization for publishers. And we're doing an Audience Development Deep Dive today actually, because this product is also

Christian Kallenberg [00:00:02]:

Welcome, dear listeners, to the new episode of the Audience Development Deep Dive by Sprylab Technologies and We Like Mags. One of my first guests on the We Like Mags podcast was Benjamin Kolb, CEO of sprylab technologies, with whom I talked about audience development and who is sitting across from me now. At the end of our conversation, Benni and I had the feeling that not everything had been said about the topic, and that's why we decided to dedicate a series special to the topic of audience development, in which we would interview guests together about the various aspects of this increasingly important topic. And, Benni, I'm really, really grateful that you, as a digital expert for publishers, are co-hosting this format with me. And tell me: What are we talking about today?  

Benjamin Kolb [00:00:53]:

Hello Christian. Our guest today is York Walterscheid, Managing Director at CeleraOne, market leader in Germany in paywalls and monetization for publishers. And we're doing an Audience Development Deep Dive today actually, because this product is also in Audience Development. Hello York.  

York Walterscheid [00:01:15]:

Hello, Benni. Hello, Christian. Very happy to be here today. Thank you for the invitation.  

Christian Kallenberg [00:01:19]:

Hello York. Tell us again in detail: What is CeleraOne's product? As the market leader, what are you particularly good at? And what can the others perhaps not yet do so well or learn from you?

York Walterscheid [00:01:29]: Yeah, what we offer is - basically you can divide it into like three modules. One is the paywall, or I would rather say a gate, the stopper for the reader, where we ask him for something, namely, for example, a registration or a subscription. That's why I always like to say it's more of a gate. Something that restricts content access. This is very closely linked to our own developed real-time audience development module, which means that our gates are always intelligently played out based on tracking, segmentation, rules, and not just simple or flat, but rather we try to convince the reader at the right time with the right approach to actually do something, for example to take out a subscription. And the last thing that belongs, if you like, to the Paid Content Suite is actually the topic of SSO, which means Identity Management Single Sign-on. We know this from the typical Facebook and Instagram worlds. Your identity is managed somewhere, and it then enables you to log in across all platforms. That's part of it.

Christian Kallenberg [00:02:44]:

Okay, what is your perception in general? Where do the publishers in Germany stand in the area of audience development?  

York Walterscheid [00:02:51]:

Very heterogeneous like the whole industry, I would say, in all areas of software. So, of course, there are somehow the big pioneers in the market, so Axel Springer is one of them, but also extremely exciting approaches at Tagesspiegel, of course at the Süddeutsche Group, the NOZ is very progressive in many areas. So it's a bit difficult to say. It's always so driven by: Who is actually in the lead there right now? Who has experience in this area? And who sees the issue of audience development as an important one? So that's why we have such an incredibly heterogeneous landscape of pioneers and publishers who are playing with data and audience development, and publishers who are not small either, but who tend to lag behind.  

Benjamin Kolb [00:03:42]:

SPRYLAB and CeleraOne actually have the common goal of improving the publisher's business model, I would say. Of course, with our publishing suite, we are also partly concerned with efficiency issues on the production side, but also with payment models for apps and websites, as well as advertising and third-party monetization. What do you think the future model for publishers will be? Where does the focus of revenue generation lie, for example? What do you see as good opportunities for publishers to improve their business model on the revenue side?  

York Walterscheid [00:04:15]:

So I would already say that there is not so the one model, but I think the important thing that publishers have to think about is actually if you want such: What revenue pillars are there actually? There is no one revenue pillar. At the beginning, everyone thought that the revenue pillar for the future would be advertising. But advertising has a lot to do with reach, and if you somehow get 0.X cents per click and have no reach, then that's not the right revenue model. A second revenue model is clearly something like the subscription, but there are also other revenue models like somehow affiliate or something that you might do in your region and can then draw benefits from it - maybe go into selling some products, selling tickets. There are brands that are so incredibly strong - not only are we currently working together with SPRYLAB on an exciting customer in Hamburg, the Hamburger Morgenpost, but it is also our customer. And they have such a strong brand. So I would say that if you walk through Hamburg and ask people, "Do you know the Hamburger Morgenpost?", I would say 90% know the Hamburger Morgenpost. So you have a very strong brand, so you should also think beyond that: What can you do besides advertising and advertising revenues and subscription revenues - what other revenue pillars are there that might be extremely exciting?  

Benjamin Kolb [00:05:38]:

What's in the product suite at CeleraOne? You said you also do audience development. Audience development is a broad field ranging from user engagement and loyalty to the optimization of monetization, churn prevention, and everything that expands or maintains the audience or, in this case, the readership. What do you do in particular? What makes you stand out in particular?  

York Walterscheid [00:06:02]:

Starting a bit broader, I think it's important to say that we at CeleraOne have said from the very beginning that we actually have to live this in the same way as in other industries for a long time, namely in this way of thinking as a marketeer and in the way of thinking as a funnel. And if you think about the funnel, there are exactly what you just said, Benni: different stages of the reader or the potential future subscriber. And along this life cycle and along the different touchpoints that you have from the very beginning - we like to call them "flybys" - they come in and out and in and out, reading on multiple sites, all the way to those that you have already converted to a subscriber and have to be rather careful that you don't lose them again, or rather that you can think about: How can you increase share-of-wallet, drive cross-selling, or up-selling? At all these points, we have always tried to consider: How can we as software support them? And that's why we've positioned our software in such a way that it can provide support in all of these areas. It starts with the flybys, who are anonymous to begin with, to segment them in some way based on their surfing behavior and then to provide them with tailored offers. This could be: "If you want to read more, register," or: "Hey, I've seen that you read a lot in such and such areas. What do you think about me putting together a newsletter for you on those areas?" Even up front there, it starts with the anonymous readers in some way, segmenting them and somehow intelligently engaging with them based on those segments. In this area of engagement up front, I think it's also exciting: We've built our own small recommendation engine, which means that the intelligence we have there can also be passed back to the content platforms that play it out, which is classically a CMS. With the CMS, if it's built dynamically - just like you build software - you can interpret the data that we're tracking and the segments that we have and display content that's accurate to the person or, let's say, to the cohort, which, of course, if you see more of what people like to read, might also drive engagement and thus increase the likelihood of having the opportunity to convert to a subscriber in the first place.  

Christian Kallenberg [00:08:26]:

Who are your contacts at the publishing houses? Are you on the technical side or is it more in marketing? How has that maybe changed over time as well?  

York Walterscheid [00:08:35]:

It has definitely changed. In the early days and in the first five years of CeleraOne, the contact persons were clearly technicians, CTOs, CIOs, rather areas with a technical background. What is now very clear in the last one, two, three years is that we have much more contact with producers or with marketeers. So it's quite clear that this whole marketing thinking and that a marketeer is also the one who works there the most on a daily basis. That is clearly changing in the industry. What has been common practice in other industries for a long time - as I said earlier - is now also becoming more and more common among all publishers. And here again, there are of course pioneers who have been doing this for a long time, or have been doing it for much longer, but you can see that there are already a lot of potential customers that I talk to who are right there. So my conversations, my sales pitches now focus much more on products and marketeers than on technicians.  

Christian Kallenberg [00:09:38]:

What is it like at SPRYLAB?

Benjamin Kolb [00:09:40]:

At SPRYLAB, we have already had a change in the field of contact persons. In the past, we were very much in the hands of the CTOs and CIOs with our products - of course, Publishing Suite. But it is becoming increasingly clear that the decisions no longer lie with them alone, but above all in the possibilities that can be offered in the area of audience development. So what options do you have to connect marketing automation, for example, or to do it yourself? The whole topic of recommendation is becoming increasingly important. So all these topics that go in the direction of marketing sales are contributing more and more to the decision-making process. And as a result, we now actually always have several contacts, if you will. Of course, because we are also involved in production, the topic of technology is also important, because it is connected to many other systems. That means we still have these contacts to discuss interfaces and so on. But the decision-making process is already moving. It's clearly going in the direction of: How can we generate more revenue with a product like this that we're introducing, and how can we get our business model off the ground better? And these kinds of questions are playing an increasingly important role.  

York Walterscheid [00:10:47]:

I'd like to go into that again very briefly, because you're right, of course. That's exactly how it is with us. In the decision-making process, the products and brand managers are increasingly coming on board, but because our software is also a component of the customer's overall architecture and is also integrated into the overall landscape, the technician is of course very important. In the context of our integration, the focus has also shifted in recent times to integrating business workshops much more strongly into the integration, because they are the ones who actually use it later. The business people have to think about what they want to sell: What products do they want to sell? What should they cost? Which segments do we actually want to address? And so on. So you can already see that. That's been a big change now, too.

Benjamin Kolb [00:11:34]:

Okay, York, let's come back to the topic of audience development. What is basically the most effective method for you? What do you do best? How far does your automation go? What are the strengths of your product, where you are perhaps also a pioneer in Germany?

York Walterscheid [00:11:50]:

We are clearly the pioneer with our product in Germany. I have to say that there is nothing that can really hold a candle to us in any form. But in the international context, there are of course some exciting competitors who are also good. But I think what we're particularly good at, because that's been a bit of the idea and philosophy of CeleraOne from the very beginning, is the whole issue of tracking data, using the data to build segments from it, applying rules to the most diverse segments, applying offers, and also checking the offers in A/B/X testing. I think that's what we're really extremely good at, and we can do it in real time. We have developed our own - we call it "in-memory" - technology, and with this in-memory technology we can of course record, process and expand all data in real time and then use it for targeting, i.e. per click and with tens of thousands of simultaneous readers. And that, I think, is just the exciting thing, that you can say, "There are ten thousand readers or ten thousand readers right now in this second or in this minute, and each of those individual readers could get a different offer based on their cohort or their segment. And the very moment he clicks, he's re-categorized again, re-divided into a segment, and then in that segment, there's just some rule going on that gives him an offer." So, and that's not somehow 2 days later that you approach him, but at that very moment when he shows interest or overriding interest in some topic. That, I think, is what we are extremely good at, and that's where the founding idea came from in the end, to develop a streaming technology of user streams and to run some kind of actions based on these user streams. And that ended up with actions for the publishing industry.  

Benjamin Kolb [00:13:48]:

How can you imagine what the impact is when you apply such a real-time technology in marketing automation? Can you show what the impact is? So, I don't know: "25% more subscribers compared to traditional products", or how do you argue that?

York Walterscheid [00:14:04]:

Exactly as it is with all software manufacturers in the area of conversion, I believe: If you can play and look, that is, if you can simply work with the software and change some parameters again and again, then based on that you can look in the reports and see if that has brought something. That's how SEO works, and that's also how conversion in the area of subscriptions works in the end. The more you can play, the more transparency you get about the ideas that you had, that you put in there, the better you can convert somehow, of course. Do we know across the board exactly how much that has done? That's a bit difficult for us, because we came into being and grew when there was basically no paid content. That is, there wasn't someone who we necessarily replaced, but there was nothing before and then there was us. I can only ever describe it in such a way that, of course, with all the measures that we have and everything that we incorporate there, we try to ensure that more transparency and more opportunities to play are given in order to set up exactly better conversions month after month, year after year. That is, what you think about is, of course, always to look within a publishing house: Have I been better now than last year? Now, of course, the Corona period is extremely exciting, and all of our customers have of course had insane increases in their surfing behavior or in their page impressions. But we've also seen that it's all somehow translated into more subscriptions. You can't say one-to-one that it's just our software, but at least the software has kept up with the whole situation and has made a positive contribution.  

Benjamin Kolb [00:15:41]:

Hopefully, the Corona period will soon be over in that the major restrictions will be over for the time being, at least in the summer, so the next exciting question is of course: Now that customers have many subscribers, what can you offer in the area of churn prevention? That is, if subscribers want to drop out again. There are a number of things on the market that go in the direction of analyzing the usage behavior of customers, subscribers, and reacting to it, or noticing when, for example, payments or subscriptions are canceled or something like that. What exactly do you offer there?

York Walterscheid [00:16:16]:

The basis of all this is, of course, our tracking and our segmentation, just as it is at the front of the funnel towards conversion, also later in the funnel. And the publishers, our customers, naturally have the option of setting up relatively simple segments that could also go in this direction. For example, they could set up a segment where they say: "All users, all readers who are subscribers and only come once a week, and when they do come, they only have a session duration of a few minutes, they are at risk for me." So, of course, each publisher has to define for themselves what they consider to be an at-risk potential leaver, but that's exactly what the publisher can then set with us and name the segment accordingly. And all users who exhibit this behavior then end up in this segment. And of course you can do a wide variety of things with this segment. For example, if you say: "They don't come on-page very often anymore, so they only visit my website once a week," then you could export this segment and the associated users of the segment to a CAM or mailing system, for example, and address them off-page with a mailing campaign. These are very clear, classic retention measures.  

The other thing, of course, is that when they do come to the site - they do come from time to time - you can of course make them an offer at that moment, i.e., play out a notification that now gets him as a return visitor with the flag: "Watch out, he could churn," something special, i.e., somehow a special offer. Or - and this is actually quite exciting, a customer of ours is currently doing this - he is thinking a bit along the lines of: What are these subscribers or these subscribers that I have gained? What are they actually doing? What are they actually reading? To understand: What is their surfing behavior? And can I derive anything from that, so that I can then make another offer based on that? So I think the exciting thing is to take a close look at which people have actually taken out a subscription. And what are they doing? And what actually drove them to do it? Was it really just Corona or was it something else? And then to actually somehow think about and develop strategies and offers that could somehow inspire these subscribers in the future.  

Christian Kallenberg [00:18:48]:

Now you've just said it: Some publishers do it themselves. What is the ratio of all the measures you just described? To what extent do the publishers come up with the good ideas themselves, and to what extent do you provide a little help and advice?

York Walterscheid [00:19:00]:

Both and. So we have the topic of consulting, business consultancy - now not only technical consultancy - that's where we come from, we are a tech company - but also this topic of business consultancy, we have now perhaps significantly strengthened in recent months and recent years, so to give much more input at the level of business consultancy. We have now also started to organize regular customer events where all customers come together and exchange ideas: "What is actually going well for you? What is going well for you? And how can we derive any common ground from this? Or what can I take back to my company as an impulse?" We're doing the same thing at the product level, in that we're now also working a bit more closely with customers there on a quarterly basis, saying, "What's actually on our roadmap?" and asking customers once again to tell them, "What are the big issues you see right now?" So we are engaging much more closely in dialog with our customers, and have also established our own Customer Success Management. In the past, this was typically just called "customer care" in the tech sector. But we actually see that we have to support the customer's success. Only if the customer is successful will he perceive us positively and stay with us for as long as possible. So the topic of customer success and consulting is a very, very important and big one.

Benjamin Kolb [00:20:19]:

What you said earlier with the: "We'll reach them again in a different way via a mailing", I found that quite exciting, because just recently the major platforms have also reacted again to how you can actually improve churn prevention, especially with subscribers, by creating possibilities that you can, even if you don't use the app at all, for example, but have taken out a subscription under iOS, that you can receive notifications when the subscription is cancelled again somehow in iTunes, so that you can then react as someone who has made the offer, so to speak, and can then also create churn prevention. Of course, there are now also great opportunities to create audience development with entry-level offers, with promotions that have been added to the subscriptions, in the sense that you can better maintain your subscriber base or, let's say, create soft opportunities to get into a subscription via entry-level subscriptions. So a lot has been done on the big platforms Google and Apple in any case with the simple payments. While you were still on the subject of Corona, were there any developments during this time, which, let's say, also started very quickly for us, where many publishers came and said, "We have to create digital offerings very quickly now"? Were there any exciting customer projects, exciting things that you can tell us about, that took place during this extremely interesting time?

York Walterscheid [00:21:35]:

Corona - which has of course accelerated the process: the fundamental digitization, the reduction in sales in the offline area and, as a result, the much stronger focus of the management of the individual publishing houses, i.e. not only of those who have always worked in the product area and marketing - they probably always wanted something - but that the resources are also released and that the management also sees: We have to do much more. So, of course, that was clearly an acceleration, that many topics were discussed anew. We also noticed that many projects were launched during this time, not just paywall or paid strategy projects, but also quite often: You are actually taking another look at your architecture. We saw that quite clearly, which was super exciting. Unfortunately, sometimes you were put off because they said, "Well, you, paywall - great, but I think we have to clean up our CMS first." So quite exciting that there are many, many topics - old legacy systems still sitting around somewhere, where they say, "Before I actually go to the topic of Paid - I have to clean up my CMS first. It can't do anything. It's static. I'm also still connected to SAP. I actually have to ask myself whether this is actually the right system for the future." You can already tell, yes. So in all areas it is being discussed again whether the system is actually the right one, whether the architecture is actually the right one. What you've also seen, of course, is that smaller and medium-sized publishers have also dedicated themselves to the topic. And that is perhaps what has changed the most. The leads that I have now received as a CeleraOne salesperson were of course in these areas. Even small and medium-sized publishers have to do something to earn money in these areas.  

Christian Kallenberg [00:23:21]:

Let's stay with the exciting customer projects, not just from the Corona area, but in the area of retention and loyalty overall. Are there any practical examples you can talk about?  

York Walterscheid [00:23:31]:

Yes, I think I've already touched on one topic here a bit in the churn topic. That is certainly what one of our customers does. To analyze it in more detail and understand it better: Who actually took out the subscription? I think that's extremely exciting. And what offers can we now make to ensure that we keep them and that they don't all jump ship again after Corona? That's one issue. The other thing that I found exciting was the whole issue of scoring. So how can you actually use all the data that you have, both our data, but perhaps also other data that you collect at the publishing house - how can you perhaps - it's called "propensity scores". That is, how high is the probability that this cohort, which is there right now, is willing to subscribe, that is, to take out a subscription? That you actually go in there again much more with data as well. There are also some of our customers who have afforded themselves teams in the direction of business intelligence. So there, too, you notice that a lot more is happening in this area of data and dealing with data in a clever way. Another customer of ours also uses exactly this kind of scoring to play out the metering model more dynamically. Metering means that you count the queries of a user, so, for example, you set one to a politics segment - whoever belongs to this politics segment and has a few other parameters, may read three articles in this area, in this segment, and then something happens. Then there is always the question: How often do you change this "then something happens"? Now you might say, three articles, but wouldn't it be much better to release four articles, because then the conversion rate increases? And of course you can manually move that back and forth every day and see what's actually better. Are two articles better or ten articles better? But there are also customers of ours who are now starting to work with such scorings, which are calculated or partly calculated themselves, but in particular calculated and then this setting of the correct value - whether it is only three articles or ten articles - is then done by an engine and no longer by ourselves.

Christian Kallenberg [00:25:49]:

Let's stay with the segments for a moment. And, Benni, maybe you can say something about that, too. This is also a topic that we discussed a bit in one of the previous episodes: The more content a user gets that is tailored to his or her interests, the greater the risk that the user will get lost in his or her information bubble. What is your view on this?

Benjamin Kolb [00:26:09]:

Exactly, that is correct. Of course, recommendation engines can be used to suggest content to end customers, who of course end up in just such a bubble. However, there are very good mechanisms to ensure that this does not happen. There are many recommender models, so you can simply suggest content that is similar to what I'm currently reading, or you can do this based on your own user history, or you can include the history of all other users, for example, to form cohorts with similar reading behavior. And then you can also mix into that a certain diversification of the content that's played out. And if you mix all these models, i.e., not only integrate your own history, but perhaps also focus more on what you are currently reading and what other customers have perhaps also read in this context, then you can actually get out of this bubble pretty quickly. And what we also do with our customers is to include a certain amount of diversification, i.e., not to include topics that might be of direct interest to the customer based on his history. You can do this very well, for example, if you often look at a recommendation. Below articles, there are four to six slots that somehow suggest articles. You can play very well with the order of placement from different models, analyze the whole thing, and then actually evaluate via AI what is actually the best click behavior. And that is not necessarily what keeps the user in the bubble. That's actually quite exciting to see, but it's often the case that customers jump off the topic. That's also the crucial thing, that you don't build up the offer in a bubble-like way.

Christian Kallenberg [00:27:41]:

York, how is that working for you? To what extent do you use AI for the paywall and the dynamization of the paywall?  

York Walterscheid [00:27:47]:

Maybe a very short point to Benni again, because I see it the same way. We have also developed our own recommendation engine simply for the reason that you have it in your product range, which is great, but we also have customers whose CMS can't do that. And that's why we had at some point based on this data that we track there - also just in the direction of conversion data we actually do it exactly so that we can give impulses to editorial systems or CMSs or content delivery systems to make just exactly such a suggestion system. And there, too, we do exactly what you said, Benni. We don't just take our own user history or similar articles, but we actually make sure that as many parameters as possible are included so that it somehow has lookalikes, i.e. has a bit of similarities, and thus also from other segments and cohorts that are a bit similar, then this bubble is less strong, because I think that is already crucial that you also put content in there that is not just always in the same cosmos. To your other question: Well, we don't use AI in that sense yet. Of course, we also calculate topics in real time with our real-time engine, and use a wide variety of statistical models to calculate scores, for example. Is that artificial intelligence? It's difficult to say. My partner and co-director, Moritz Hilger, who also founded the company, would say that this is not artificial intelligence, but statistics. I believe that if you were to look at a definition of artificial intelligence, there's a bit more to it than that, which is why I would say that our current status is that we use a wide variety of statistical models to calculate precisely this kind of thing. And, of course, this is done fully automatically. There is no one sitting there calculating it. We play around with it. Is it artificial intelligence now? Do we use a lot of external data sources that also go into it in order to expand this entire data head? No. We mostly do this on the publisher's data. Perhaps especially in the discussion around GDPR, i.e. DSGVO, and the topics of first-party data, I think it's important to say: each of our customers has its own setup, which means that when we work with scoring, it's always on the data of the respective customer. We don't do this across several customer systems, but it really only runs in there. But, of course, AI, artificial intelligence, dynamization is also a very, very big topic for us, which we always include in all our discussions.  

Benjamin Kolb [00:30:32]:

And can you talk a little bit about your current state of research on that? So what I could imagine is that you could use advanced statistical methods or maybe even artificial intelligence to make price determinations if you want to have dynamic pricing, to do what you said earlier: From how many articles and perhaps for which topic is it actually best to bring up a paywall for which customer group in the first place? There are a lot of parameters and adjusting screws where, I think, as a human being, you probably won't find the optimal solution right away, because there are simply far too many parameters in the whole system, but where, of course, if you have enough users - you always have to add that data plays a decisive role here - you could actually build wonderful models that learn.

York Walterscheid [00:31:18]:

Yes, of course, it's quite clear. I'm right there with you, Benni, because a human being can't make such profound decisions very, very cleanly in this area, because of course a machine can process so much data in milliseconds, no human being can do that. That's why the question is always: Where do the decisions of those who work with the software actually come from? It doesn't matter whether it's in your keyword search or similar areas or ours. How does someone know which metering is the right one or which segment is actually the right one? Or which paywall offers are actually the right ones? That's why I believe that the machine, AI, or basically statistics models like that can help tremendously. You can also start off a little softer, by saying: At the beginning, there's a suggestion system. It doesn't have to be a machine that you can't stop again, or that works in an uncontrolled way without knowing exactly what's happening. But I think the future clearly lies there, because you can't incorporate so much data into your decision that the decision is really well-founded. What we did there: A few years ago, we did an extremely exciting project on the topic of dynamic paywall. The question is: Which article is actually paid and which article is free? If you have such a classic barrier of freemium, there is always the question: What do you actually put on paid and what not? We ran a pilot with Google News Initiative and a publisher in Luxembourg that was extremely successful. We went through a wide variety of individual steps, starting, of course, with the question: How large is the cohort that we actually get, on which we can then run our tests? So this "N=", what is that actually? And then it was really a matter of saying: The editor decides what is Plus or what is Premium, versus the machine decides. And that was such a longer collaboration, a longer pilot at that level. In the beginning, it's more of a random decision. The machine just says 50/50, this one it's paid and that one it's not, and decides it that way until the end, where you then let a lot more parameters flow into the decision of what's actually paid. And of course you can scale up these parameters as much as you like. And the more parameters are included, the better the decision will be in some form or other, and the less such a decision can actually be triggered by a human being, because to include all these parameters in this one decision is actually not even conceivable and not even feasible in terms of time.  

Benjamin Kolb [00:33:58]:

You've talked a lot about the dynamic paywall, i.e., when the paywall or the gate, as you called it earlier, comes up, but it's also exciting to see what the final price will be. You could also make this dynamic, i.e., offer certain segments, target groups, different prices for the goods that they have to purchase.  

York Walterscheid [00:34:19]:

Absolutely, yes, the Dynamic Paywall project was a bit longer ago. So far, we haven't been able to establish it on the market the way we would have liked to. In fact, a customer recently approached us and we are working with them on dynamic pricing. We'll see what comes out of it, but it's all about how you can actually show different prices to different segments. And what effects does that actually have on conversion? I'm very curious about that. We're working flat out on it right now, and I assume that the first version, version 1, will be released somewhere after the summer, when we'll see what's actually so exciting about dynamic pricing.  

Christian Kallenberg [00:35:01]:

Of course, it's hard on the self-confidence of any journalist - I'm allowed to say that because I'm one myself - when you no longer know what's best for your readers. What are your experiences? And how do you perhaps also deal with fears from editorial offices?  

York Walterscheid [00:35:14]:

It's a big topic, of course. It's been with me for a few years now, the topic of artificial intelligence in the software area of publishing houses, whether it's production software or monetization software. The reservations about a machine are huge. So no matter where we come up with such topics - and this has not only been the case with this Google News initiative, but also before I had such topics that then immediately it is said: "Okay, that's robo-journalism. Or it's the machine. I no longer have any control over what it does and whether it actually does it really well." Of course, the editor, because he has been doing it for a long time, has an extremely good feeling, but in the meantime, the surfing behavior, the data that exists, is so manifold that, I believe, a machine is absolutely helpful at least in the impulse "I'll help you with it," because no one can process so quickly 365 days, 24 hours a day, never sick, is always there and can process this data further and make a suggestion system or also make the decisions. That's why I think that's exactly the right direction to go in. And there are a lot of reservations, which is a bit of a shame, because you can see that the openness in the industry is also limited - not with everyone, of course - because either there's the dead-beat robo-journalism in the content area and we have similar ones. In the area of monetization, there are also similar dead-beat arguments, which is why it's a bit of a shame that openness is so - not at all, as I said, but already at some publishers - very restrained, instead of running tests or pilots or daring to get involved. There are other areas, other industries, that are already testing a great deal and are having good success with it.

Benjamin Kolb [00:37:15]:

Yes, what I find particularly unfortunate about the whole story is that with such automated systems, or even now in the use of AI in certain areas, so much is being taken away from what is actually not the task of journalism in the true sense, i.e. as far as distribution is concerned or, let's say, finding search engine keywords that fit best and constantly optimizing them, or setting links between articles. So that's what we already had in the previous episode, that digitalization has basically added so many tasks that journalism is fading further and further into the background. It's hard to understand why this isn't being approached more openly, why all these burdens aren't being removed by a machine. Now we are also noticing that the golden egg in the publishing industry has not yet been found in the sense that every publisher knows how best to monetize - especially in digital, of course. In the digital sector, they often want to try things out. And especially in audience development, the most exciting part is of course to try things out and be flexible in what we do with our software, so that you can let off steam as flexibly as possible on the front end and also try out business models. In the paywall area, it's probably mainly metering, the offer page. What marketeers always like to do when they want to try things out are A/B tests, or preferably of course A/B/C/D/E tests. Automation may also play a role here. What is your state of the art or what you can offer on the market?  

York Walterscheid [00:38:44]:

We do indeed offer A/B/N testing options in the Offer Pages area. Unfortunately, these are not yet dynamized, i.e. the suggestion system or even the machine decides what offer pages or what offers are actually the right thing for this segment, but today it is still very simple to set up our cockpit, then of course to play out A/B/C/D offer pages there, and of course to look at the performance of the individual pages: What actually worked and on which path do I then continue? That works, of course, but the topic of dynamization is also on the agenda - unfortunately, this often fails due to lack of capacity. So of course that's an investment you have to make, and we'd like to make it, but it has to fit in with our overall business model in some way, but of course we can see that. When we talk about innovations and opportunities for further development, these topics of dynamizing offer pages are extremely exciting, as is the fact that data-based AI models or statistical models actually suggest segments. So why does the publisher have to click together the segments himself, which is easy, but how does he actually know that this is a good segment or a bad segment?

Christian Kallenberg [00:40:08]:

And is there someone in the publishing house who can do that? Not every publishing house has the capacity to build something like that, to build knowledge.

York Walterscheid [00:40:17]:

Exactly, so building up knowledge. Right, of course we tell all our customers: "Be careful, you're not buying a simple plug-in here, which drives up a wall somewhere and then someone converts or not, but you're buying conversion software. That's why I don't necessarily see myself as a paywall, but as conversion software. And if you bring such a powerful software into your house, it must be clear to everyone that you also need resources for it. So someone has to operate it and play with it. If you don't want that, then I would advise everyone to use a static paywall, because you have too many options for that. But, yes, you can tell. Sometimes the capacities aren't there or the knowledge isn't there, but there, too, it's exciting - because you asked earlier, Christian - there, too, you notice that publishers are also hiring people from other industries who come precisely from this audience development topic. So we're seeing that in our industry as well, that more and more "career changers" are coming into the industry, who may have learned somewhere else, learned in the e-commerce business, and are now coming over to the publishing industry and actually bringing their e-commerce knowledge into the mix. That's super exciting.

Benjamin Kolb [00:41:32]:

But not every publisher can, let's say, build up the knowledge themselves and make a long-term commitment to such a topic. Do you also have a solution for publishers where you say: "We'll do the consulting for you, we'll create everything for you, and we'll just make sure that your sales go up"? Or do you have partners who do that for you?

York Walterscheid [00:41:47]:

I already touched on this a bit earlier. We created an area some time ago - called "Customer Success" - and that's exactly where we want to build up this business consulting. We now have colleagues who deal with this very issue and help our customers on a consulting level, i.e. really business consulting. This can go so far that we meet regularly and actually talk about segments, about the data that we have tracked, about the possibilities of playout and then of course also get impulses from our experience with all our customers and let them flow in. So, do we have a customer who says, "Why don't you do full service and set it all up for me"? We don't have that. It's not going to work that way either. I think people realize that in the area of conversion - and it already starts with SEO - they need someone to deal with it. And now the question is always: How much do you deal with it? But then the SEO man or woman does half and half, so half SEO and half paid content. So someone has to take care of it, yes.

Christian Kallenberg [00:42:56]:

It's a blatant cut now. It's become standard practice that we seem to only have guests here who are extremely fit, not only on the computer but also in the kitchen. York, you said in the preliminary interview that you, too, are at least an enthusiastic cook. We already shared recipes with our conversation partner last time. Any quick thoughts on Corona-friendly family dinners or anything like that? Now don't say delivery service.

York Walterscheid [00:43:19]:

Well, it's become a bit more like Lieferando, to be honest, or a delivery service - it doesn't really matter who. Because when you spend a whole day homeschooling and working from home, and then somehow still have to motivate yourself to cook in the evening - sometimes you're just exhausted and use the app to order food somehow. But, yes, still like to cook. Extremely like to cook with friends, of course. That's, I think, what's nice about it then. It's suffered a bit, but my daughter says that the best saltimbocca still comes from me, so cooking is cool.

Benjamin Kolb [00:43:57]:

Good, York, thank you very much. I found it super exciting.  

York Walterscheid [00:44:00]:

I'd love to, I'd love to. It was fun, thank you.  

Christian Kallenberg [00:44:03]:

And if you, dear listeners, have feedback on this episode, let us know. Did you like it? What would you do differently? Want to talk to us yourself? Send us an email at podcast@welikemags.com. Thanks for listening and see you next time.

‍