Friday, March 6, 2009

What is the real proof of service reliability?

We all are experiencing tough times. That’s why every step you take in business now can bring 10 times worse and deplorable results in case you make the wrong decision than ever before.

If talk about web-based services, what are the actual criteria of vendors’ reliability?

Really, how the common user can define if the service is trustworthy and what are the key factors?

There are 5 tips listed in TrackVia blog post by Ed Dunigan in order to clear things up what the customer should take into account and what makes the service reliable.

I tried to briefly sum it up:
  1. Great back-up services
  2. Test all features
  3. Challenge the Service Department
  4. Read the Manual
  5. Ask to talk to current drivers

I really doubt those are the reasons to consider the vendor reliable and its enough to make right decision. Let’s see exactly why.

Firstly, I can hardly recall the vendor that doesn’t provide data back up in present time. This is not the issue to worry about. If the company goes down the customer will get the data back in any case. The other thing is that nobody cares what you gonna do with this data afterwards and how to migrate it somewhere else. So, transition plan is completely the users concern yet.

Secondly, don’t waste your precious time on testing ALL features the system offers. It’s better to spend it on checking if the service fits your specifics, trying to set up only the functionality you need right now.

And what I definitely wouldn’t recommend to do is to bother the Service Department asking them useless questions in order to try their knowledge.

Concerning Manual reading I’ll tell you this: if trying the service there is a need to constantly read the Manual, this is definitely not the solution you are looking for.

Well, there are many companies out there that setisfy all above listed points. There is a live example of the Coghead's failed venture that complied to the Ed's list pretty well. Did it stand to the Five Star Crash Test?

So, what is that than? I believe if the company is capable to stay afloat because of its own revenue it’s already something you can trust. Investors money is not something to rely on, especially now.

Before making any decisions, check everything properly.


  1. I certainly would require the ability to perform point-in-time recovery of any SaaS application. User initiated data deletion is by far our most common recovery problem, and being able to re-create a complex system at the point prior to the logical failure is essential for any app that serves more than a handful of users. You cannot have an error by UserA force you into recovering an entire system from last nights backup and thereby cause the loss of all data by UserB, UserC ... User999999 from last night until now.

    That usually means recovering data to a parallel system in a new spot and merging the recovered data with the data that way modified since the 'incident'. We maintain the capability of doing that on all our applications/systems.

    From what I see on the news, that capability should not be assumed to already exist on SaaS services.

  2. Hi Jane,

    We'd love to have your readers test out our service department - it won't be a bother. We put the "service" in software-as-a-service, and I think Ed is right, that customer support is a key indication of a company's reliability.

    Thanks for the great posts on SaaS.

    Chris Basham, CEO

  3. Hi Jane,

    thank you for a really insightful post, and it's hard to disagree with your point.
    Sure, the main concern would be company's profitability and no dependency on the outside financing, such as VC. In these days companies that are depending on a financing rather than on their own revenue are really vulnerable because it's up to their investors to decide whether to shut their doors right now or to wait a little longer and burn some more venture capital..
    While it is easy to gather the information about public companies (only to see their stocks heading south), there is no much you can learn about private companies.
    Some of the private companies may be really stable if they have proportional infrastructure, no debt to repay and no investors to depend on, but here customers are again on their own, as no official data is available.

    I also agree that the platform should speak for itself, and only a hands-on experience would matter.
    Free trial users definitely need to test customer support effectiveness, opening a few support requests and asking for assistance. However, I don't think that pushing support team to their limits without a real need is a good idea. Customer support team may only be efficient if they concentrate their efforts on their real-life customer needs. If the team has plenty of time to spend, then it is probably over-expanded, which is not a good sign, especially now.

    If, instead of testing useless features and trying to push the vendor to their limits, you concentrate on your real-life project, you should be able to get it done in a few days (or even hours!). You may start you production work while in the free trial mode. This way you may even get a "free lunch", even though Ed Dunigan doesn't believe that a "free lunch" ever exists.
    "Free lunch" does exist, but only if you concentrate on getting the project done and if you are working with vendor's support team really effectively.

    ...Waiting for your next post - this blog is addictive ;-)

    Val Karmazin