[Sigia-l] SUMMARY: Estimating content analysis & migration

daniela meleo danielameleo at yahoo.com
Tue Jul 13 23:04:32 EDT 2004


Here is the summary of replies I received for how to
estimate content analysis & migration tasks.  Thanks
once again to everyone who chimed in. 
The folks who were asking for my help were VERY happy
for the advice & guidance.

... Daniela Meleo
_____________________________________________________

<< daniela meleo (danielameleo_at_yahoo.com) >>

I'm helping estimate some of the tasks for a planned
content migration / content management project.
I've  checked the wiki and this list, but couldn't
find much specifically on estimating.  

There are about 3000 pages of content (don't ask me
how long a "page" is...of course I don't yet have that
level of detail! ) 

The tasks in scope include analysing content and
determining content types for input to a CMS; 
determining the level of content "cleansing" required 
to migrate from the previous system to the new
templates; liaising with subject matter experts to
review & approve final content; importing the content 
into the new CMS; defining the workflow process and
display rules etc for the content. The site is being
revamped to better handle new and changed product info
 to help customers locate, choose and order products. 

I haven't had any hands-on experience in a project
like this -- any tips on where to start & expected
productivity or useful estimating metrics for this
sort of work would be most helpful. 

thanks.. 
===== 
Daniela Meleo 
IBM Australia, BCS User Experience 

_____________________________________________________

<<John Howe (john_at_wyriwyg.com) >>

Daniela, 
You might find some pertinent info. in Chap. 11
("Staffing a CMS") of Bob Boiko's humongous "Content
Management Bible." 
See also pps. 701-8 of that book re "Staff Modeling."
And Amy Warner [www.lexonomy.com] gave some useful
info. on resource allocation 
for content inventories in her workshop at the 2003 IA
Summit in Portland: "Advanced Taxonomy and Metadata
Design." 
That may (?) still be on www.asis.org or elsewhere
online, or available directly from her. 
- John 
_____________________________________________________

<<Eryk Orlowski (e.orlowski_at_argonauts360.pl)  >>

i would: 
- after determining the cleansing level, think of data
format in new system, 
- determine requirements for data format transform, 
- design some engine for batch data processing. 

the rest is quite easy - if You have any project
management background, You will estimate the costs
using work breakdown structure.
If You have no idea  and don't want to have one, hire
some project manager, that will prepare an estimate
for You, using project management 
and software engineering standards. If this is within
Your scope of interests, will give You some key
vocabulary to search for. 

_____________________________________________________

<<Hilary Marsh (hilary_at_contentcompany.biz) >>

Hi Daniela,

I am actually working with IBM in the U.S. on a
project just like this one. It is a combination
content/IA effort, joining the taxonomy with the
content strategy analysis. (I'll give you more details
privately, if you'd like.)

The level of effort for migration varies depending on
the CMS you choose, where the content is currently
housed and the skills of your internal technical team.
I've seen it take place in one day, and I've led
projects where it took four people approx. 5 weeks to
migrate 1200 pages of content.

The content analysis and typing activities are
intensive but fun -- I can definitely give you
insights into this, as well as defining workflow and
display rules.

____________________________________________________

<< Jonathan Baker-Bates 
(Jonathan.Baker-Bates at oyster.com)  >>

Hi Daniela,

I don't know whether the following is helpful in your
case, but we recently used a somewhat experimental
approach to some of the content issues you describe. 
It turned out to be pretty good: allowing us to build
a content inventory as a by-product of the work of
re-designing a site's IA and graphic design, while not
 having to worry too much about the CMS. The client
wanted us to use the "same content" (spread across
various content  types) but obviously we needed to
audit and re-jig this so that it would suit the new
design.   So, we used a Freehand document to record
which page "modules" appeared on each page template;
an Excel spreadsheet then listed all pages on the site
(and also captured the taxonomy we wanted them to
conform to) and  recorded which templates and modules
appeared on them. Each page was then  produced as an
XML file via a schema which described the properties
of the module designs (e.g. a module would contain a
title, text, drop-down menu and  one or more links).
The resulting pages were then editable with MS Word 
(2003) and populated with content by editors using a
web-based version control system. The resulting XML
was then be transformed into whatever format  was
required (e.g. HTML, PDF layouts, SQL statements,
etc.). True, this  stage was time-consuming, being
essentially a copy/paste fest, but we had  enough
people on the job for it to go reasonably quickly. The
Word files could also be passed around over email, or
even printed out and filled in by those  that were so
inclined!

More importantly from your point of view perhaps, the
system allowed us to quantify exactly what content was
needed quite early on in the process. 
For example, by the time we had the first cut of the
taxonomy and the spreadsheet, we knew that we were
going to have to generate 200 summaries of existing
documents, and could tell exactly what new content
needed to be created in other areas, even to the
approximate word count etc. Over time, we could also
get some rough management information about the
content generation process - for example the system
could tell us that 35% of all the content needed for a
certain area of the site had been loaded into the Word
files. The PM liked that aspect of it! Word also
allowed us to validate input to a certain extent (e.g.
throwing an error if the user hadn't put some required
content in, or had exceeded a word limit, etc.)

It was a bit more complicated than that, and I'm going
from memory, but if you want I can give you some
further details. 
Jonathan
_____________________________________________________

<<Lynn A Madsen  >>

Daniela,
This is exactly what I am doing right now for American
Express. The first thing you must do is a complete
content inventory, and engage all of the content
owners. Supply them with the inventory and let them
validate the accuracy, relevancy and timeliness of the
content.  Make it clear to them just because content
exists now does not mean it gets migrated to the new
CMS. You must put together a governance model for both
decommissioning and creating new content, and that is
where you will identify the workflow process.  You can
also introduce the content types in the governance
documentation. 
That's a high level.  Good luck.
_____________________________________________________

<<Kathleen Klein (kathleen at kleininfodesign.com)  >>
<snip>
In general for estimating copy cleansing, I ask the
client for a representative sampling of the content
and estimate based on that how much clean-up will
take. Having the client pick content, you spare
yourself from having to do a lot of analysis just to
figure out what is representative. - Kathleen
_____________________________________________________

<< Sarah Rice (rice at seneb.com) >>
Daniela:
I've worked on a project like this. I found that doing
a thorough content inventory to be immensly helpful.
There's a sample spreadsheet up on the AIfIA Tools
page that I submitted which could serve as an example.
It's a great place to start. 

The project I did was a migration as well as a
redesign, so I did an initial inventory of all pages
that I thought would get migrated, then I morphed that
information as the information structure changed. The
best part of the spreadsheet was that it captured
information about EVERY PAGE that was going to get
migrated. Filling out that spreadsheet helped to
determine all the questions that needed to get asked.
It was a tedious thing to do to fill it out, but it
uncovered about 90% of the things that we needed to
address in order to do the migration. People
questioned why we needed to account for EVERY PAGE. I
found that each page seemed to be unique (even if it
was the content and the metadata that was associated
with each page). All that information would eventually
get used, and gathering it up front helped get
EVERYTHING out in the open before migration began. It
really helped to have a realistic view of the scope of
the project. Of course, it was overwhelming, because
it was a lot of content. It's almost like people
preferred NOT knowing the scope of the project, just
so it would seem managible.

No one on the team had done a project like that as
well, so we spend significant time figuring out where
we were and then figuring out what needed to happen
next. We really couldn't see too far down the road to
plan ahead.

Here are some tactical suggestions:
1. start an initial content inventory first. It will
give you the "lay of the land" and show you what
you'll be working with. It will also reveal other
things that you will need to figure out.
2. for workflow, it was best to document initially the
CURRENT process that people are using. Then, assess
that and figure out which parts can be automated by
the system and which ones must be covered by
organizational process.
3. determining a list of content types: I gathered
suggestions from everyone, compiled a master list.
Refined the list (you'll find that you need to define
"what is a 'content type'"). Have others review and
augment. You'll have a list to begin with. Then, it
may change as the project progresses.

I don't know how big your team is, but I worked on
such a project for 9 months, and that didn't include
actual migration of content. I was the only IA, and I
found that 1 wasn't enough. 3 would have been better.
I spent some of that time training others to be Jr.
IA's to complete the work. I think the total page
count was about 5K.

Hope this gives you some idea. I'm happy to answer
other questions you have. I know I would have really
liked to have someone that had walked the road before
me to talk with from time to time during the project.

Good luck!

Sarah



		
__________________________________
Do you Yahoo!?
Yahoo! Mail - 50x more storage than other providers!
http://promotions.yahoo.com/new_mail



More information about the Sigia-l mailing list