So lately I’ve been seeing a new class of web site that leans more towards custom development than a horizontal portal. The factors influencing a custom development approach mainly fall into the proliferation of Web/RESTful services and better front end tooling in AngularJS, Bootstrap, and similar items. Both Forrester and Gartner have labeled horizontal portals dead although not as explicitly as I just stated. With continuing evolution of best practices, we have to question the place of horizontal portals with Bootstrap hitting it on one side and modern web content management systems like Adobe and Sitecore hitting them on the other side.
I’m not surprised to see that both Forrester and Gartner are pushing “portal is dead” and there is some truth to it. The reality is that a portal fills only the web site and mobile web channel and you still need a bunch of api’s to fill the other channels and do it as cheaply as possible. It’s also true that every front end guy out there loves Bootstrap, AungularJS, etc. Perficient’s eXperience Design guys are no exception. To some extent, the added flexibility in the UI justifies it. Perficient sees it as multiple clients in a healthcare, financial services, and a variety of other industries.
Where Integration / Cool UI Tools May Not Work
All that said, these approaches solve some problems while being incredibly sexy at the same time. However, they do not solve the following:
- True integration of content. This is web content although documents can easily be included. It’s hard. Portals spent a lot of time and money solving it and they do a good job of integrating content while allowing contextual search and personalization.
- Easy addition of pages and functionality. Those who love the internet of things will say this is irrelevant and in the consumer world that has a smidgen of truth. In the b2b or intranet world, it’s not true. People using laptops or desktops still think in terms of pages. Many companies still want a flexibility when adding or modifying pages. In contrast, the custom approach pretty much demands a developer for even minor changes.
- Surfacing and/or integrating with existing applications…….where no api exists. In a perfect world, every app should have api. In reality, too few have a pi’s. Bootstrap with api’s doesn’t solve this.
- Intranets – intranets have links, content, applications, sso, etc. In the case of an intranet which has to solve and service multiple use cases, I don’t see this approach working. I do see specific applications within an intranet using the Integration / Bootstrap approach though.
- Very flexible partner and customer portals. When you have to bring reports in document format and chart format, include apps to do a variety of lookups and reporting, integrate to knowledge management systems, and make the search work across it all then a custom approach may prove too costly.
- Danger of tightly coupled applications. I can give multiple client examples where custom development started out great but became a morass of code over 4-5 years. That morass meant tight coupling which meant longer development cycles …….. which raised costs. Despite the somewhat “heavy” nature of horizontal portal development, they do push you towards a loosely coupled development approach.
I cannot gainsay any company using a custom development approach.. In reality, they are building something that mostly an app and only partially like a portal. With the added UI flexibility and ease of use of these new libraries and UI’s, it makes sense to do that. Portal technology will lose these types of web sites more and more often.
So here’s what I see in the next five years:
- The horizontal portal market will shrink. Every web site that is very app like will head towards the api/cool ui driven model.
- Portal vendors like IBM, Microsoft, and Liferay will respond and build in support for the cool UI’s into their themes…….without breaking the content integration and portlet model. It will be hard but they will do it if they want to remain viable options.
- A lot of companies will flirt with the api driven model and in the case of very complex sites that aren’t just apps, they will fail and that then return to portal or at least a set of social networking tools.
- Cloud vendors like Salesforce are going to take advantage of their inherent strengths to to launch these type of app driven self-service sites quicker and the api driven model may not matter as much. Sales and Service Cloud integration to Communities is already great……. Even though no cool UI model exists like BoostStrap or Angular. Of course, with Force.com or Heroku, you could use SalesForce as a pure api provider.
Portal is not dead………. But it’s not heading into a huge growth curve. The maturation of integration services and the introduction/use of UI tooling will shave off a set of sites previously in the horizontal portal realm. Frankly, it will make business sense to do that. But this development approach will not fulfill all the use cases for a variety of internal and external sites. Horizontal portals will remain viable but they better continue to evolve.