Too Candid?

Javelin Research published a great report today on Online Account Opening which covers the market, the business case, the value proposition, and ten different vendors, all in a lot of detail. We don't have distribution rights so I can't give away any of their conclusions, but I will say that Andera was portrayed very favorably. If you can swing the price tag for the report, I'd highly recommend making the investment.

One area we got knocked on was our occasional growing pains that I have described on this blog. At the risk of getting in trouble for quoting the report, Javelin said:

Kroll is admirably forthright for acknowledging Andera’s problems, but the growing pains are a cautionary note in an industry that demands reliability.

After reading this, it occured to me that after being very candid about our problems, I should take the opportunity to also be very candid about how we've addressed them and the significant progress we've made.

There have really been two trouble spots for us as we've scaled our company:

System Capacity. As I chronicled here, issues occurred in February, as we saw an unprecedented and unexpected spike in transaction volume that we didn't properly plan for. We now believe the spike was due mostly to macroeconomic factors related to an increase in the national savings rate, that caused people to pull money out of securities (stocks, bonds, etc) and into cash, in the form of savings and money market accounts. Because of this surge in volume, our system occasionally slowed to an unnacceptable speed at peak periods (in some cases 14 second page response times) and even came down a couple of times over a couple of days.

As I chronicled here, we immediately reprioritized our ops and engineering teams, and challenged them with completing an already aggressive full year of upgrades in three months. The first of these upgrades was rolled out within a week, and with it came a material improvement in response time, along with a complete restoration of system stability (no more outages). More upgrades were rolled out along the way which continued to shave multiple seconds off response time, and last week we were back down to sub-five-second second page responses (at 95% percentile), which is within our target range. As of now, we have two major upgrades left (one related to our database and another related to our load balancing architecture) to round out our plan, both of which will further improve things. We believe we are today in a comfortable place, and given our new capacity planning and scaling framework, well positioned to tackle any level of higher volumes going forward.

Implementation Capacity. More than a year ago, in January of 2008, we found ourselves overcommitted as a result of faster-than-expected sales. Sounds like a good problem to have, but that didn't make it any less hard to solve. In blog posts here and here, I outlined the issues and our plan to resolve them, which among other things, included doubling staff on the implementation team and overhauling our implementation process from the ground up. It took some time for the changes to take effect, and with all the team growth there continued to be some slippage as people were coming up to speed and learning curves were being overcome. But by the middle of 2008, things were clicking and the shackles came off as we saw monthly launch count double from five to ten during the summer, more than enough at that point to handle new clients and deplete the backlog. Since then, things have continued to accelerate, but we have handled the volume well given our new team structure and improved processes, so there really has not been an implementation capacity issue for quite some time. Additionally, the new Editions we announced recently will further improve things, as many of our new customers going forward will benefit from much more streamlined setups, as evidenced by our recent two week implementation for a Premium Edition customer.

All of this is not nearly to say "mission accomplished," since things will continue to evolve and the challenges will change. But I am saying that I feel like we are quite well positioned to handle future growth now, unlike where we were a year or more ago. This is also not to say that we are perfect today - we're not, and like any company that operates in the real world, things still slip through the cracks and we do make mistakes. But fortunately these instances are manageable and relatively infrequent, and I think most importantly, we recognize the importance of minimizing them.

So am I wrong for being so candid on this blog? I hope not - I know it will turn some people off to find out we have issues, but I hope it convinces a much larger group, which recognizes that every company has issues, that ours are under control, taken seriously, handled competently, and less bad than most other companies out there.