The BIG Blog

Gartner Sets Record Straight on SaaS BI on Demand

February 26, 2009 10:13 pm Esmart Analyst Reports, SaaS, on Demand BI

One of the main difficulties most businesses face when selecting software is separating the hype from the reality. This also applies to considering web-based on demand business intelligence analytics.

Gartner found that of the five commonly held assumptions – three are entirely false and the other two only partly true.

Businesses often don’t appreciate how quickly software can be improved after the initial launch. With such a competitive market, software development vendors are pressured to get products to market ASAP, and no longer wait for perfection in terms of either feature richness of functional performance. However, with feedback from early adopters, development continues vigorously and the applications can be almost transformed in the first 6-12 months. I think this is actually a really good process, as it helps to validate exactly what users do want, and not waste money developing performance or features that are not appreciated.

Of the five most commonly held assumptions about SaaS BI models:

1. SaaS is less expensive than on-premises software – TRUE

SaaS applications do not require large capital investment by businesses for licenses or support infrastructure. This can significantly reduce the total cost of ownership over the first two years. After this time, client site deployed applications can become more economical in terms of financial reporting [amortisation impact], however this is not necessarily true for an operational IT expense perspective.

2. SaaS is faster to implement than on-premises software – HALF TRUE

Speed of implementation for SaaS is faster for simple applications, however one must deliniate between initial implementation time and the additional time taken to deploy it to all users, which can take 2-3 times longer. As the complexity of processes and integration increase, the gap decreases. This is due to the larger percentage of the deployment time spent on customization, configuration and integration which can be equally difficult for both models.

3. SaaS is priced as a utility model, similar to electric companies – FALSE

Many vendors claim to charge on a usage basis, but in most cases they must commit to an agreed estimated usage independent of actual use.

4. SaaS does not integrate with on-premises applications and/or data sources – FALSE

Companies can integrate web based data using either batch synchronization, real-time integration using Web services or at the user-interface level through mashups.

5. SaaS is only for simple, basic requirements – FALSE

Whilst there are limits to customisation on SaaS models, the feature set of many applications rivals that of on-premise versions. SaaS vendors provide development platforms that enable high levels of configuration and the metadata level. The area of tightest constraint is in end-to-end processes requiring complexworkflow or business process management capabilities.

I hope this gives more confidence to those businesses considering SaaS models. For more on SaaS BI and vendor options available.