Morpheus : You take the blue pill, the story ends. You wake upand believe…whatever you want to believe. You take the redpill…..you stay in wonderland…and I show you just how
deep the rabbit hole goes. The Matrix
Does the portal concept still make sense on the Internet? Nope, read on, stay in wonderland and I’ll tell you why.
Lets restate what a portal is supposed to be: A website that offered structured pathways into the web, offering a range of services and content (many their own, while others would be covered by partners). These portals would be the gateway through which most internet users would pass to the rest of the web. And when companies like Yahoo! and Lycos had their IPO’s the market agreed.
For companies like Yahoo!, Lycos and MSN being internet portals have caused endless problems. They have been at sixes and sevens, trying to be all things to all people, fighting with their own internal subdivions over the use of an overarching brand, style, and even navigation. All a direct result of being portals.
They have been heamoraging users and not growing their reach except in key services like instant messaging and email where they got an early head start. But they are loosing ground to the MySpace’s and Google’s of this world. In fact, most of their organisational and strategic problems are related to them being portals.
Portals are tied to an old fashioned view of the media. A view that did not get that with the internet power shifts away from large top down media companies. Companies that were set up to produce fairly general content to be consumed by large groups of consumers.
In reality we are moving deeper into a world where every person can become a producer and not just a user of media. This is a concept that companies like Google got from the start.
If you accept this vision of the changing media world – a world where it becomes easier and easier for anyone to create a website, post a comment, and upload a video – it follows that on the Internet there will be a great variation of websites, and opinions, and also lots of competition for people’s attention.
In this world entry points to the web are very valuable. If you control these entry points you become the gatekeeper to this fragmented media landscape. Which is exactly why Netscape – the company that blasted the first hot air into the first net bubble – was thought to be so valuable. It was thought that through control of an internet browser, one could become the gatekeeper to the web’s content. Following the same logic, the concept of the internet portal at that time did make sense.
And today, with mobile phone portals this arguments still holds, but just. Mobile operators have a huge captive audience, its cumbersome to type on the mobile interface and therefore to search for services or information is difficult to do.
Yes, until search engines evolved and especially when they evolved to rank listings according to quality, the links on a portal was significant. But search ranked by quality, introduced by Google’s Pagerank system changed the landscape dramatically. (Google ranks a page according to how many pages link to it. Thus if web users link to your page, Google reckons it merits a higher value.)
An even earlier way of how search engines rank websites were the building blocks for this change. By analyzing how popular a site is for a certain set of keywords the trend towards focused and simplyfied pages became significant. If your website does lots of things – like most portals do – and list them all on your home page, the home page ranking will be diluted.
Pagerank will be diminished on a jack-of-all-trades page, because the homepage will give no clear indication of what it’s about, and the web users won’t recommend it by linking to it.
This is because it will feature many different keywords. This coupled with the fact that websites are unlikely to link to a page which is just a collection of different services, is deadly. No single Portal could possibly list all the websites a user likes to use. Users’ needs are too specific and various. And thus the websites run by users tend to link to services directly.
Another way to understand systems like Pagerank is how they value the links outward from a page. A link from a high value page will transfer high Pagerank value to the page it links to. But the value to be transferred is shared, so by having three links from a page, the value would also be split three ways.
Thus with a traditional portal’s home page with lots of links to various of its services, the result will be that its links’ power to transfer value will be diluted. The more links they have the more true it will be. The value transfered from a link on the home page of Yahoo! to its email service will be shared with its links to its other many services.
Search engines’ indexing hierachy favour and therefore create a web where there are lots of specialised “champion” websites. Often the characteristics of an internet service further enhances this “champion” effect because of so called economies of scale (Ebay) or a critical mass of users (think IM, MySpace).
The result of all this is that search engines have become quality selection engines. Users have come to know that used well, a search engine gives them access to the best of breed web content and services. Once you follow this logic you can see how the old style portal concept is really dead. A concept for those that won’t experience wonderland.
What compounds this problem for the portals even more is that users are increasingly becoming so used to the mechanics of search, and are increasingly using it as a way of navigation. (I for instance, when I want to look up information on Uganda’s dictator Idi Amin on Wikipedia, just search for “wikipedia idi amin” via Google and click on the first link Google serves up. I’m there. Why bother entering the Wikipedia address into the browser and then searching for Idi?)
Search as navigation is becoming an unstoppable force.
The portals are squeezed on two sides. They are not specialist sites – the sites that tend to be recommended by search engines – and users don’t need portals to point them to these sites.
Time for a bit of futurerology: Users will navigate the web in two ways. By searching via keywords and by searching via brands (or rather sub-brands and sub-sub-brands). In other words when searching for something and you don’t have a brand to mind – that you know will address your need – you will just do a normal keyword search.
Otherwise you will Google/ navigate directly to the sub-brand you know and trust to work well. (I for instance use Yahoo! Mail and I Google “mail yahoo” to directly go to mail.yahoo.com.)
And its already happening. The majority of Yahoo! mail users do exactly that and never see the Yahoo! front page. Even though Yahoo! tries to persuade us all the time to go and have a look and circulate its traffic.
Yahoo! is arguably the most successful portal in that unlike MSN – they actually have one or two channels (Yahoo! travel – as opposed to the large staple portal services: search, email, IM) that are a relative success.
But even Yahoo! has been realizing that users come to it for specific good products and not primarily because they are a portal. That’s is why Yahoo has bought a huge one act pony Flickr (a photo sharing community site) but wont be changing Flickr’s name. Nor have they been adding links from the Yahoo! home page to Flickr. It does not need to. Flickr is a top ranked website already thank you very much. It would be stupid to change its name because a recongnisable brand is paramount if you want users to find (search for) a service.
The so-called Yahoo! home page is actually just a glorified site map. Their users increasingly go directly to the services Yahoo! offer bypassing the front page.The successful Yahoo! channels and services are a success because they are good, not because they are on the Yahoo! homepage.
People have been wondering why Google does not link to its other products from its home page. It’s exactly this reason. They don’t need to, and if they did it would reduce the simplicity and usability of the Google.com page, making it less popular with users.
So what are portals to do? (To be continued.)