We are now a few months into our redesign project and it’s time to reveal some of the findings. For those of you who read the previous three posts about the dysfunctional ad environment, the broken site experience and the blueprint for our then future web sites, some of this will be a repetition, but we’ll keep that part short and try to get you some interesting numbers.
From the beginning, we set out to create sites that were better for the reader and the advertiser at the same time. Doing this means handling a few different time perspectives at once – so every change we’ve done in the user interface, ”the product,” we’ve done in order to reach a change in user behavior and at the same time enhance the user experience.
This may sound like it’s a given, but as we talked about in the previous posts, it’s really not. Instead, a lot of media site web design and development has gone for short-term gains in user behavior that runs the risk of creating bad user experiences.
This is all fine in a market with constant growth, where the absolute number of available ”readers” keeps increasing. But we’re nearing the end of that continuous growth and it will be more about share of market than absolute size. This in turn means that the overall user experience will become much more important. So the strategic portion of every decision becomes more crucial.
Let me give you a quick example. Doing a slideshow of images instead of publishing them all on one page will probably give you a direct hike in page views – so the user behavior will change in a positive way. But the user experience will be negatively affected by this – and in the long term perhaps prevent the user from coming back.
Winning the battle but losing the war – that has been the de facto philosophy of many design and product changes, but we can’t see it because the war still rages on.
The reader perspective – a little less looking, a little more reading
Our focus was to create better sites for the readers and better sites for the advertisers. And the way we measure this – the user behavior – is by looking at some really basic metrics and some more unusual ones. So how has it gone so far? Well, thanks for asking and quite excellent, actually.
Combining these numbers for the last four weeks before the migration to the new sites with the last four weeks gives us some interesting numbers.
Total page views, a metric we were prepared to see go down with the redesign, are up by 27%. Unique visitors per week are up 14% on average and visits per week are up on average 23%.
These reach numbers reflect a lot of things, of course – how popular the content is, how good we’ve been at marketing ourselves. But the page view numbers are also an indicator of how good the site functions, and their growth is outperforming the visit numbers’ growth by a fair bit, not least because of a superior performance on mobile.
In essence, people are finding more reasons to come back to the sites and are also finding more stuff to read while they’re there.
The comeback of the article
Since we operate sites with both blog content – our expert bloggers, blog stars and celebrity bloggers – and regular editorial material, we have the opportunity to compare these two types of content. This is interesting since the blog pages have changed a lot less from a UI perspective than the editorial content pages have, so it’s a great indicator of the success of the redesign in itself.
We’ve seen that the article reading has outgrown the total reading increase by a lot. On the biggest site of these three, mama.nu, the growth in page views for articles has been 140% since the migration, which means that that type of content now makes up a larger part of the total page views. We attribute this to three things: the changes we’ve made on the start page, the fact that we’re now properly repeating the flow of start page content at the bottom of each article page and the reading experience for the articles.
Another statistic supporting the success of the redesign is that on that same site, mama.nu, the bounce rate from the start page has decreased dramatically – by 43%. This tells us that the readers are finding a lot more to read when they come to the site directly and that they’re finding it more quickly, since the time spent on the start page also has gone down by 43%. The other two sites show similar decreases of bounce rate as mama.nu, something that we’re properly proud of. And, the readers spend more time consuming content than looking for it – exactly the way it’s supposed to be.
The advertiser perspective – quality inventory galore
All in all, this means great results for the reader. But what has happened from an ad perspective? Our goal was to create a higher in-screen rate for ads – to deliver more ads that will actually be seen. With the new grid, we dramatically decreased the number of potential ad slots on each page and we placed these ad slots so that two ads never should be on the screen at once.
We measure in-screen performance for all ad impressions and that’s where the real story is. (An impression is when the ad is loaded from the ad server. The tools then measure if it has been displayed on the screen – according to IAB standards – and thus visible for the reader.)
One ”homemade” KPI – called ad yield – we’re looking at is the number of visible ads per page view, where the maximum, of course, is the number of ad slots. It does not take into account fill rate, so this is purely a measurement of how good the page functions as an ad environment – not how attractive it is on the market or how good your sales department is.
For us, it was very much an issue of increasing the number of solid ad slots, with the right format – the 980 panorama, which on the display market is really the only viable format and on mobile it’s the full-width equivalent. Comparing the same time periods as we used for page performance, we have grown ”sellable” inventory (ad yield) on these three sites by 44%. That’s a huge number for us, and I think it would be for any publisher. And this with sticking to our principles, full-width ads and one ad per screen.
Another thing we’re observing, though it’s much harder to quantify, is that we’re seeing a lot better creative in the ads. The agencies have really understood that these ad slots are great and that their creative really gets the chance to shine.
So, in short, so far we’re really happy that we’ve been able to do what we set out to do: Create sites that are better for readers and advertisers. Because that’s what the numbers tell us, and we know that the changes in user behavior that we’re seeing are firmly rooted in an understanding of the user experience, so we’re not worried that they are a fad. They are, actually, fact. And that’s how we drive our development process – based on facts, data and knowledge.
While releasing more sites on the new grid, we’re constantly working on refining what we have. We’re building a state-of-the-art A/B test environment which will allow us to run even more tests. Even if we think the numbers are good now, we want them to be better and better still.
With that in mind, expect more follow-ups, more facts and figures and even more detail going forward.