uk-netmarketing Archive (2011-2015)

Hi Suzy,



Sorry, to come back to your question and share a bit of detail with you.



CMS platform wouldn’t be my first choice for a big technical build project
(like the Olympics site).



FWIW here’s how I would go about it;



I would recommend first identifying the high level stakeholder requirements
and scenarios,  from business, admin and end user perspective   Keep it
high level -  as detail requirement docs become  redundant as soon as the
team knows more,  which they will on day 2 of the project.  I would look to
save time and sanity not writing reams of requirement documentation.



There are formats for doing that type of thing,  here at ESW towers we
follow agile methodology and use stakeholder and story format – it is fast
and focusses on maximising development time to produce working software
early.   The site is then improved iteratively in short development
cycles.   We prefer that to writing big documents that go out of date as
soon as they are written.



Prince 2 would be suitable for a development where a lot of stakeholders
who have no experience of a complex build project within a big
organisation,  it creates a good framework and ensures documentation,
checks and sign offs are in place at the right levels.   But Prince 2  is
comparatively  heavy weight and expensive  –  but if adhered to will
increase chance of success.



I would then set about identifying the right development team who have
experienced in the same (event) sector with the same scale of
development,    by the sounds of things this is not a design / marketing
agency who have a contractor who knows a little bit about drupal or
sitecore job  (=disaster).



The right team will probably have preferred platform(s)  for the CMS and
ecommerce on which the web build will happen.   There are a number of
suitable platforms for this,  so I wouldn’t get hung up on it.    However
licensing costs,   and access to the source code under license would be a
key determinant for me,  if you wanted to take the development in-house in
the future say.    The team would be able to actually demonstrate
customisations that are similar to your business requirements  (not just
hypothesise about them and present them in pretty mock-ups).    They will
also be able to demonstrate experience of integrating the other
applications  that will be required to meet the project requirements.
Big projects normally have multiple integration points so demonstrable
experience would be important to me.   On the subject of integrations the
supporting systems should have well formed APIs and good documentation –
again,   demonstration of integrations would be important.



Construct of the team would probably look something like;   business
analyst, account manager, technical project manager (SCRUM master),  system
 architect,  front-end web designer /  UI designer, developer,  database
admin, sysadmin and QA tester.    One person may perform multiple roles,
  but a big project like the one you describe is way beyond single
contractor / web designer.



Ideally that team would be experienced in a development process (like
agile,  SCRUM &/or XP) ,  they would be adept in story estimation and
(importantly)  tracking project velocity – *velocity tracking  is key*.
They will be able to show you project  management systems for development
management,  QA testing,  build testing (development / staging / live),
support and issue tracking.     They also have a testing framework and use
unit testing and automated QA testing,  like selenium.



I would definitely look for an agile team  (it saves so much money and
risk!)  but would be very mindful to find a demonstrable track record,
‘agile’ is easy to say,   but can be a disaster if done badly as you remove
most of the traditional project constraints.     Can they shoe you velocity
sprint to sprint on a live development?  Burndown?  Do they know what
planning poker is?  That kind of thing…



The team I would look for would also be able to load test your application,
and will know about things like funkload and jmeter  -  that way they can
test what happens when increasing numbers of people do things concurrently
on your site.   That prevents embarrassing and expensive massive system
failure, and things like that.     They’ll know about mitigating that kind
of risk,  things like caching and load balancing.  To those ends I’d look
for established ISP and hardware vendor relationships with suitable SLAs in
place.



In an ideal world the team would also be able to use proper analytics and
MVT tools,   more importantly they will be able to help setup the right
reports with the right level of detail to make them useful.



Code and build management are also important, and code deployment tools, as
is a clear product roadmap and an idea of technical debt and legacy
management.



I’ve probably missed a few important things – but hopefully that should
help you with asking the right questions of the right kind of company.



regards



a*n*gus



PS If anyone shows you  a Microsoft Project Gantt chart, run a mile.





--
Angus Phillipson
Director
http://www.WORKSsitebuilder.com <http://www.workssitebuilder.com/>

T: +44 (0)208 780 6350
M: +44 (0)7710 438 972
angus at workssitebuilder.com
http://www.linkedin.com/in/angusp
http://twitter.com/AngusPhillipson
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mm.chinwag.com/pipermail/uk-netmarketing/attachments/20111125/901d20ea/attachment.htm