So, since my last blog post on this subject Oracle has acquired ATG and hybris has acquired iCongo. Success in the marketplace along with significant functional improvements to the platform has allowed hybris to become a genuine competitor to Oracle ATG. The technology stack up better and in terms of doing business, who would you prefer to contract with?
ATG does not properly support B2B (we have worked on a number of heavily customized ATG B2B sites. You have to build the functionality from the ground up. 50% of hybris implementations are in fact B2B and they have a B2B module which provides significant sums of functionality straight out of the box and does not rely on a significantly customised version of the B2C platform.
Multi site, multi tenant and internationalisation. There is virtually nothing in ATG, you have to build the lot or take customised components from 3rd party integrators to achieve the same thing. Hybris’s key strength is in this area and is fundamentally driven off the PIM and CMS capabilities. We are working on a project right now for a leading UK apparel retailer that has multiple brands, with half a dozen international sites for each brand going live at the same time. Development effort on our accelerator – under 40 days including skinning to achieve this. You could not write the technical design for this on ATG in 40 days.
Somebody above was extolling the benefits of ATG CMS, Wow, this is very very wrong. Content Management (CA) in ATG is not a PIM and is only a content editing and deployment mechanism. It is usually front-ended with a Stibo, Fatwire or other content management solutions, which means a messy integration of two content management systems. It does not publish content to any source other than the website product catalog. It is also difficult to work with as it uses “switching data sources” for online and offline catalogs. A common problem with this design is catalogs that are often out of sync causing failed deployments and an IT re-baseline. On hybris we have simply not come across this problem. I’ve seen hybris blow ATG out of the water on CMS capability alone in evaluations (functional demo stage) and again this is one of the really big benefits of hybris over ATG.
Although websphere and weblogic are supported the de facto implementation for the past 4-5 years has been JBoss. Now Oracle is advocating weblogic (they own it) which is more complex than JBoss and has a bigger footprint. It will cause massive headaches on projects around configuration of the application to run efficiently on another app server.
Content deployments invalidate the cache degrading the performance of the website Customers are encouraged to update content only during off-peak traffic hours. So you spend millions on the licenses and they tell you this afterwards – more needs to be made of these issues as it’s just stunning what ATG customers have to put up with.
ATG’s commerce reference store (CRS) used to accelerate a project is not production ready and in fact scored (using open source PMD rules) as one of the poorest code bases compared against ATG customer implementations. Unlike Portaltech’s project accelerator or the hybris accelerator it should not be used as the foundation of a production site.
In a managed services (hosted) environment business users have no way to modify or create scenarios (personalization rules). This is managed by a separate UI (ACC) which provides direct edit capability to the website. Managed services locks this down for SLA purposes. The business needs to submit a written request to the managed services to change promotions/scenarios. Is that for real I hear you ask? ATG could never get SaaS right and I’ve even less faith in Oracle managing to.
Implementations are often lengthy (12 months or more) and expensive. They will be hard pressed to provide positive implementation references. Wal-Mart (Sam’s Club) took 2+ years to implement plus another 9+ month refactoring project to fix all the issues. Ann Taylor was a direct project by ATG and was 9+ months late with several hundred thousand of cost overruns. El Corte Ingles took nearly three years to implement and Tesco is already two years in with quite literally hundreds of millions spent and no live site to shoe for it. I wonder how many senior heads rolled at Tesco over this?
ATG knowledge, portal, response management (email) and self-service will no longer be sold by Oracle. Bang goes one of the key differentiates to hybris at this time. (hybris plan to release this functionality soon).
In fact what is Oracle’s strategy for ATG? Overall the platform was built in 1990s, there have been release changes but no major re-write. This makes the platform have a large footprint and difficult to work with. Do they re-write, if so what happens to the old ATG customer base, surely they won’t re-build and pay Oracle for new licenses? What call centre or CMS will they use etc. I genuinely believe that ATG will simply die a slow and painful death – we are already seeing this in the marketplace. What deals have they won in the last 12 months in the UK for example? The mass exodus of staff from services, support and sales surely won’t help their cause and will exasperate the implementation issues even further as system integration partners also begin to look at other technologies such as hybris.
The one saving grace is that ATG will now have a decent search engine through the Endeca acquisition – one can only imagine the horror of the license costs now associated with acquiring ATG: weblogic licenses, ATG licenses, Fatwire licenses, Stibo licenses, Endeca licenses, RightNow licenses. What an ungodly creature ATG has become.