|
|
The following are projects that I've done under the banner
of Westslope
Software, the company I founded January of 2001.
Background.
For this project, Dell was undertaking building a shared cart implementation between their Dell Online website and their
inside sales call center application. Some colleagues of mine at Seilevel took on the project, along with members of Dell's
technical staff and some additional contracting technical talent.
The Project. I was brought in to help with managing the
development project. We hit the ground with 2 weeks of planning,
and followed with 6 weeks of development leading into User Acceptance
Test. My job was to assess the risks up front, propose risk steps
to mitigate those risks, and to organize the development team, and
finally to help them execute against the plan.
The main risks were dependencies on other projects, a large scope,
and tight timeline. We did our best to stretch the timeline, from
4 weeks to 6 for development. We limited the scope in some critical
areas, freeing up resources to lock down other parts of the project,
and eliminating some key bottlenecks to delivery. Finally, we were
able to secure commitments for early delivery dates on some key
dependencies. On the other hand, we had to develop alternate plans
for certain functionality rather than depend on those other systems.
Overall, the goal was to reduce the risk profile of the project,
with a key focus on external risks.
We organized the development team into a UI team and an architecture
team. The architecture team being responsible for delivering the
meat of the functionality, while the UI team was responsible for
all the UI interaction requirements and tying it all together. We
also brought in key resources with expertise on the applications
we were integrating (Dell Online and the inside sales call center),
as well as expert consultants with knowledge of the new technologies
we were deploying (e.g. .NET). Successful delivery of the project
enabled my colleagues to land another profitable contract with Dell,
and allowed Dell to achieve a small savings per transaction (which
on Dell's scale, means a lot of money).
Reynolds has been a Trilogy customer almost as long as I worked
at Trilogy. In fact, I worked with Reynolds back in 1995 for about
6 months. This year, they asked me back to help them with the rollout
of a new major productline to their salesforce. As the expert on
configuration, and SalesBUILDER in particular, it was my primary
responsibility to see to it that the configuration model met the
new requirements. However, my responsibilities also included educating
other members of the team about the SalesBUILDER API, and building
a new C++ dll to handle some of the more demanding mathematical
calculations used in Reynolds' models.
The architecture of this application is not what one would call
groundbreaking. It is a Visual Basic 6 application. C++ dll's are
used in various ways to support the application, along with OCX's
from third party vendors, and some custom-built OCX's as well. SalesBUILDER
is written in C++ and in a single-user mode has the same basic architecture
as it did in 1994 (the new internet-ready SalesBUILDER was an effort
kicked off in 2001, while I was the development manager of the configuration
team at Trilogy). The database is replicated to the client, and
the application runs fully on the client. This was cutting edge
architecture in 1995 and 1996. However, a project like this is worth
doing because it has ROI for the customer. For Reynolds, this application
will educate and enable their salesforce to sell a complicated new
flagship productline with a vast array of options available, many
of which are different from previous productlines.
When I talked with Zilliant, they had a very near-term need for
a highly qualified software engineer to work on custom development
for Zilliant customers. Leveraging my network of skilled professionals,
I placed a colleague with Zilliant who had 4.5 years of excellent
experience under his belt, along with the personality professionalism
to blend in with the team. Only one week separated the contract
signing and the offer being made to the colleague I proposed. Two
weeks from contract to start date.
TestComp has a great business going with two of my old friends
from Trilogy running things. First, there is Jeff Hotz, who runs
the US side of the business. Then, there is Ashish Shah, who runs
the India side of the business (SNS). Together, they provide excellent
testing services to their customers. In many cases it pays to have
a quality technical lead on site to spearhead the effort. This is
particularly true when you are stress-testing an enterprise software
application.
When Jeff asked me to help out on a tight four-week project, I
couldn't refuse. I managed the project for TestComp and their customer,
Smith Co, in Houston. We did a load testing and performance analysis
that resulted in an 80-page analysis document, in addition to graphs
and charts and other data to help Smith Co in the future. In this
project, I coordinated work with the team in India and with our
customer in Houston. It helps to have great people to work with
on both ends, as I did, but it also takes an understanding of how
to take advantage of the time differential, and cost differentials.
If I can assist you with performance or load analysis of your software
applications, or help you coordinate with a team in India, contact
me.
|