ubuntu-us-ne Ubuntu in the Non-Profit Office
Matthew Platte
51 at triopticon.com
Sat Dec 9 23:25:56 GMT 2006
On Dec 9, 2006, at 3:11 PM, Dave Thacker wrote:
> What's the biggest need of social service agencies?
> Chris Holt thinks he knows.
>
> http://www.freesoftwaremagazine.com/blogs/free_software_for_non-
> profits
Abstract: In the human services network there are many service
providers but only a handful of funders. The trend is towards
standardized reports, pulled from the central database. Once a
service provider grows enough to win a few grants, their data
management toolset will likely be determined by the funders'
requirements.
My experience is pretty much limited to a few local human service
agencies. My point of view is somewhat less than universally
applicable.
One of the growing trends amongst the funders (USDA, HUD, United Way,
etc.) is that in compensation for making fewer dollars available they
are tightening their reporting requirements. For some (Head Start)
the funder has mandated the use of a particular case management and
reporting tool for quite some time. For others, they've been asking
for more detail and analysis in the periodic reports and have
recently or will soon be mandating a common client management
platform to be utilized by the many agencies that they fund. United
Way has been pushing ServicePoint pretty heavily recently. A few
years ago the Lutherans were developing a similar, comprehensive
project; don't know what became of that one.
ServicePoint(SP) is the package I'm most familiar with. SP has been
around Lincoln for several years now. It's developed by some outfit
in Louisiana and there have been a series of federal grants feeding
the development of SP just here in our little part of the planet.
I'm sure that well over a million federal bucks have been spent
directly or indirectly on SP at UNL.
Part of the rationale for SP is that all agencies providing services
to a family or individual will work from, and build, a common shared
database. Some data is generally available, some bits are not -- and
each agency evaluates what they can or can not share with the other
agencies. The database is what you'd expect in that there's an
initial building of the family record then agencies and programs
within agencies hang their program-specific data fields on that
structure.
A benefit for the service provider agency is that the burden (and
expense) of generating the periodic report for the funders is pushed
down the org chart, out of the project administrator's office and
into the data entry cube farm. For example, at the state level, an
administrator in Nebraska Health and Human Services can pull
demographics and other reports directly from ServicePoint without
waiting for the individual agencies' mid-level managers to submit
their inconsistent and poorly-formatted paper reports. The agency
that needs to track just a little bit more about their clients is
encouraged to add that to their SP collection of data entry forms.
Indeed, ServicePoint refers to itself as a Management Information
System.
Yeah, well that's the some of the promise. The reality of
ServicePoint is that it is complicated. Complicated and slow to work
with. Slow, unresponsive software leads to bad data: duplicate
family records bloom because it's a PITA to look for the numerous
possible (mis-)spellings of a client's name. It is, of course, a
relational database so the un-structured and semi-structured data
that really tells the story of any given client is difficult to
pigeon-hole. We can count how many 47-year-old Hispanics that were
born of Asian and Native American parents but there's not a very
effective way to capture the reasons why a person requests the
services of the agency. The classic design problem of posing
multiple-choice questions because the answers will be easier to count.
If ServicePoint were to publish an API, there would be development
opportunities; perhaps they have already done so; I haven't seriously
looked or asked.
--
-------^.^--
Matthew Platte
51 at triopticon.com
More information about the Ubuntu-us-ne
mailing list