Saas is increasing in popularity – and I am no exception in succumbing to its Siren song wiles. Here is a poll as SaaS from pollcode.com which is adapting the Google “let ads pay the freight” model of SaaS . I get to setup at poolcode.com the poll shown below and pollcode.com does all the backoom pollkeeping and display work. Readers are encouraged to register their vote on the most important trade-offs in using SaaS.
However in using this SaaS service I encountered some problems that were enumerated in my very list above. For example, rollcode only allows 60 characters per observation. So I had intended to ask/define some elaborate questions/trade-offs but I had to cutback. The first trade-off was to read:
Surrendering control over a system operations, uptime, backup+ recovery and scalability versus having fairly deterministic costs tied to service level agreements
Now this is a fair degree of change required to accommodate meeting the pollcode.coms poll requirements. To meet rollcodes operational requirements I have to give up the exact phrasing I wanted in these questions. But this is not unheard of as I have talked with colleagues who have had to fit into SalesForce.coms CRM features and requirements. even with its extensive APEX library of addon capabilities. However, even shops using the inhouse ASP model for Accounting or HR or Supply Chain Management or whatever with Oracle, SAP or other ERP package suppliers have had to adjust to loss of control over IT system.
However, I would like to raise two other concerns I have about SaaS that did not quite fit the pollcode model. The first concern is related to control over a system but this is centered on functional/feature control. The question is: How many customizations and how deep must they be to compensate for loss of control of the overall design, rate of change and future direction of a system? I obviously have encountered this – I would like to allow for bigger questions, changed formatting, and some piechart graphics. But I have small control over this.
The second problem is associated with the online primarily SaaS model. This website has gone down twice in the last two months with significant outages of a day or more. Now ICDsoft, my hoster, has a pretty immaculate record of 99.99% up time for most of its servers. But I got unlucky …. Same thing could happen for your online SaaS provider. Or they could be subject to Denial of Service or other hack attacks. Or the vital routers and Internet infrastructure could go out for any number of reasons. And as more systems go SaaS a downstream critical software service provider could go down. In short, I am saying that SaaS opens up by its vary online nature a wider vulnerability window/profile than in-house server based systems.
So I am looking for SaaS/ASP providers that offer duplicated remote backup operations and more local offline emergency operational capabilities. If the system goes down I can fire up a local server that runs the SaaS software coupled with my backup data. Even in a small publishing operations my systems are so dependent on daily access I have duplicated all my websites locally so I can continue to operate if my host provider goes down.
Yes I know Ephraim and the analyst community are enthusiatic about SaaS . And as the trade-off list above implies, there are some attractive but also potentially more risky trade-offs. Be a Boy Scout – Be prepared! Oh and please take the poll above.