Dennis Mortensen is CEO of Visual Revenue, a predictive analytics company for online publishing.
AdExchanger: What has been the big highlight for Visual Revenue in the past year?
DM: We haven't pivoted into anything else so the vision with which we set up the company a year ago is exactly the same today.
In the space that we're in now, where we spend all our time speaking to editors, the thing that matters the most is NOT an increased revenue opportunity or lift. They want to make sure that if they're at the FT (Financial Times), the FT stays the FT and doesn't turn into the New York Post, or vice versa.
So what we needed to do was not come out of the gate with the message: "We can increase your views by 30 percent if you're willing to take some of our recommendations." What we say today is that "I can make sure that you can do your job faster, better, and as accurate. And, while you do that, you’ll end up having a lift on top of it. Isn't that nice?” That's certainly something we've learned. Focus on the publication, and make sure that they stay true to it. We're not here to change it.
We're not Accenture or McKinsey. We help you do what you do better.
How predictive can you get in news – especially considering your company’s services?
Let me start by answering this in [a roundabout] way - folks in retail earn a whole lot more than people in publishing. And data which retailers have available for a set of products which they sell - which could be that pair of jeans you're wearing right now - those jeans were the same two months ago, they're the same today, they're going to be the same two months from now. The past is very similar, if not equivalent, to the future and that provides retail a great opportunity.
In news, there is no past. The events in Egypt happened [for the first time] two hours ago, for example. You can certainly learn something from the past, but it's not as accurate as you might see in other industries. What you need to have in place is a learning mechanism. When something is published in news, you can learn extremely rapidly. This is the complete opposite of what you would see in retail where you might run a test where we start Monday, run it for 10 hours, come to conclusions and run with them going forward. The average article lifespan on some of these stories is less than 10 hours.
We spent a lot of time trying to come up with “if my initial prediction is slightly off, how can I self‑correct?”
Does Visual Revenue try to understand the audience or is it all about traffic numbers?
This is not necessarily how we go promote the platform.
Two decades ago when editors sat down in the afternoon and decided what would go on the front page of the printed product, you would have a set of section editors who would go pitch what they believed should be on the front page. They would have a one-hour meeting and decide then. That's a fair setup if you have a product that comes out once a day. Online though, you update the front page 50 times a day or more, and you can't have that conversation. What you're missing from that conversation is the voice of the audience. It’s not how we promote it, but if you think about it, we're the voice of the audience. That doesn't mean that we set the agenda. We just want to participate in that dialogue.
Is there a next step to predictive analytics for the publisher where it gets married to buying an audience or where they can arbitrage audience?
I certainly see it. There might be opportunities appearing for publishers to take advantage of some arbitrage window that has been created, too.
In the long run, you will have to rely on automation in some way, shape, or form. And, that makes sense because this is content-driven. There's no “winning forever” in this market, because it's connected to the content. Some story or proposition might warrant a set of opportunities over a week or over two hours. You need to roll with the content, and that's extremely exciting.
How does search (Google, Bing, etal.) and the newsroom play together? Can search be optimized for an editor who wants to optimize their front page?
Here's my philosophy. News, in its most pure definition, is non‑searchable ‑ because if I can search for it, it's not news. I don't think editors should care about search at all when they write their articles. Once a specific piece of content is thrown into an archive, somebody else should take over -somebody who's there to make sure that you maximize the value of 15 years of good content and that you get something in return from it. Don't write a headline on the day [you publish] with search in mind because it doesn't matter.
I know that's not an aggressive attitude, but that's how I look at it. Yes, I am biased. But it really makes sense if you think about it. And, when I talk about search, I'm talking about traditional search, not Google news, which is a news aggregator.
Isn’t there an important feedback loop that should occur between the archive and the editor that could help in providing some predictions on what content provides the most bang for the buck?
First of all, in this handover, you should optimize the title of the abstract, potentially even some of the keywords, the setup, apply the right HTML, make sure that the linking internally is done on some topic modeling. There are all kinds of organic search tactics which should be applied once it arrives in the archive.
Once that is done, you will make some revenue, which is small, on the individual story, but large in aggregate because you have 15 years worth of archive. And, there are strategic measures that you can put in place when you try to figure out what to write about. Some people write entirely for the archive, they don't have current content.
If you go look at Demand Media, they write for the archive and for a set of keywords. I do think that's the only value that connects both ends here.
If you're one of those publishers where 90 percent of your revenue is attached to current events and 10 percent is to the archive, shouldn't you just invest accordingly?
Let me bring you back to the most recent news at Visual Revenue. You've got an all‑star lineup of investors to the tune of 1.7 million dollars, and that's not a whole lot of money in the ad tech start-up world. Why raise so little?
The previous ventures that we've done have all been completely bootstrapped to the point where we didn't take 10 cents in funding - that attitude of owning your company outright and being able to take extremely rapid and aggressive decisions. They might be risky, but you can take them. There is no board meeting. There's a Diet Coke, a look out the window, and then there's the decision. That's the process. I am fond of that. Some people, somehow, equal funding with a success point. I'm not sure I would necessarily agree here.
Funding provides you opportunities, but you should take the minimal amount needed. I don't believe in the "Take as much as you can when the opportunity is there," because you don't get it for free. We've done this three times over before. For us, it's taking the smallest possible slice of equity that we can, and exchanging that for dollars.
You've got specific partners here in your funding announcement ‑ any method to the madness, if you will?
There are two reasons. One, we have a fantastic data background ourselves ‑‑15 years in data, three companies, all acquired, the last one by Yahoo. There's no doubt really that we understand data. One of the partners, which is Lerer Ventures, we certainly picked and asked to participate again in this round. They were eager to do so. Because of their editorial background, we wanted that Huffington Post story ‑‑ which Ken Lerer co‑founded and Eric Hippeau led himself ‑‑ to be affiliated with us, because we don't have that story. So that's a great connect for us. Then, there's the ‑ for lack of any better way of defining it ‑‑ big data stories that Roger from IA Ventures does well. We would like to kind of solidify who we are and what we do, and we do that well with Roger. Given the opportunities for us to choose here, we think we've made a wise choice in that.
Where is the Visual Revenue product going next - ads?
Where you'll see us go is not in the direction of optimizing the ad inventory. We would like to stay on the content-only side. There's a whole host of companies on the ad optimization side ‑ some of them good, some of them great, some of them not so good. We want to make sure that we are the best friend of the editor. If we cross that line, that dialogue could become slightly blurry. We don't even want them to think about whether we're really here to help them on making good, proper editorial decisions ‑‑ that they're in agreement with ‑‑ or we're just here to make another buck. So, I don't think you'll see us cross that line.
Secondly, there's the idea of us being able to utilize our platform and technology in other industries. If we get to a point where we feel we've exhausted this industry, you take a step to the left and look at other industries. We're not there yet. We would like to do one thing and do it well. As an entrepreneur, I'm fond of doing exactly that.
The third point is where you would see us expand: envision that Bloomberg terminal of the newsroom - which is the same industry. It's all about content. And whenever you have something that is slightly gut‑driven, such as "What should I write about?" you might go look at Google search trends, what's trending on Twitter, what's popular on Facebook. What do you write about? That's somewhat of a gut feel with limited data.
I want to be able to predict what to write about, what's within the universe that your reader's interests are interested, what you've written about already, and here is a set of suggestions and their values.
We help editors program the home page or the section front page. That tends to be about half of all the articles views and that's why we've chosen them. But ask yourself, "How do you think they program Twitter, Facebook, their RSS feed, or their email newsletter?" Probably with not too much science and we can add a lot to that equation. That's where you'll see our platform expand. They walk into the newsroom of the AP and you won't see 300 people looking out the window. You'll have 300 people with a terminal, each. That's where I want to take it.
Any milestones ahead you can share about the company?
A lot of people will have, just like we have, an idea of where they're going and what their mission is and we speak about this “Bloomberg terminal of the newsroom” all the time. That's how we got started and that's how we continue. We're really strict to that mission and want to continue down that avenue, and execute on it. That's what we're going to do, yes. Did we reach the 100th customer? No, we got halfway there. Are we going to surrender? Not at all. Do we see ourselves as being successful? Extremely, because what we set out to do is actually working. Customers actually like it and are willing to pay for it. We've been able to attract what we believe are some fantastic investors. We doubled our team since you and I spoke. We'll probably double it again when I speak again with you in a year from now. So you know what? That story that we've started, we just want to finish it. I don't want you to walk into Wired magazine, People, the AP or the FT, and not see one of those VR terminals, because you will.
By John Ebbert