So the RFP process is a standard (but inefficient) way of doing business. What if it was done a different way?

One of the more significant decisions you make in your life (if you have children) is where you will send them to school. It is not a decision you make lightly as it will have a major influence on how your child grows up in the world. In the UK the government provides data (league tables) but this can only tell you so much; there is more to education that the statistics tell (which are historical and do not necessarily reflect the current reality your child is going to face). You will probably ask around - seek the wisdom of the crowd. Undoubtedly the community can identify good schools and bad schools. But the best judge of a school is to go there, to look around, to meet the teachers, to see the children. Do you trust the leadership of the head? Would you be happy for this person to teach your child, would you like your child to play in this playground, (and more importantly) grow up with these children?

So why not apply this thinking when looking for a supplier to build you an application? At the end of the day, projects succeed on personalities and relationships. Will the vendor get on with the buyer? The RFP tells you little about that. What if the RFP process was like seeking a school for your child? What if you had a project open day where you welcomed suppliers in, got to meet them, and maybe even got them to compete against each other.

What if you had three intense days when the business, IT and prospective invited suppliers come together to define the project and complete against each other in teams to come up with the “best” solution.

What if you provide the suppliers with details of what you are looking to achieve and request a basic qualifier - company details, profitability etc (the stuff that goes on every RFP) and a list of clients they have built similar products for (not exceeding one page of A4 per client). And for costings you ask them to provide you with their proposed rate card.

What if you then invite all suppliers to a large venue with a space for everyone to gather, and break out areas for the individual suppliers to work in. You start with background and presentations from the business and from IT. You tell the story of what you want, the vision, a description of the current technology, constraints, assumptions, known risks, integration points, etc. You provide some initial direction as a large group, but then breakout into supplier teams, interspersing each team with your people - from IT and the business. You provide technology (access to your systems, whatever is needed) and domain expertise. What happens next is up to the suppliers. They then have two days to impress.

What if at the end of each day each supplier presents their output to the whole group. The following morning you outline what you like of the outputs and ask the teams to take that as input to work on. Then at the end of the last day each vendor puts in an anonymous sealed envelope with their estimate (resources required to build the application). Can this triangulation technique be any less accurate than the estimate given on the back of several pages specification in an RPP?

If we accept that IT projects are about people, implemented by people, then the benefit of this approach is that you get to work with the supplier and experience the relationship first hand, rather than through documents and practiced PowerPoint presentations. And for the supplier it reduces the time taken to respond and will be more enjoyable for those involved. After all, don’t people prefer to do rather than write about what they do?

A practical example of this in the bricks and morter world can be seen at this Sandwich bar in Hong Kong.

Clearly modelled on the western way of buying sandwiches, the counter layout supports the customer selecting the product (sandwiches and fillings on display) moving on to the cashier at the end of the counter to pay.

This isn’t the way things are done in Hong Kong where money comes first before the product. “Please place your order at the cashier”… before dwelling in front of the display cabinet. This results is congestion around the cashier counter and poor workflow and a slow and tedious customer experience.

Sandwich bar or internet offering, consider cultural differences before transferring the concept and content.

BA: “Certainly.  BA 28. Midnight Thirty Five.”

And that was it.  No friendly warning, so close to doing this again.

If I ask for a flight on Thursday, common sense suggests that the fact that the plane is scheduled to depart 35 minutes into Thursday, and I need to be at the airport at 10pm on Wednesday, then it really isn’t a Thursday flight.  Not being explicit and clear with this is rude business.

Except for a few exceptions, software is all about a user interacting with a system in order to accomplish a goal.  Trying to describe that premise in terms of features and functionality is fundamentally flawed as it will be nested in the language of technical implementation, not in the user interaction.  “Find” becomes “search”, “buy” becomes “shopping cart”, “check” becomes “validation” and so on.  Issue #2. The desired outcomes become lost in a smog of technical jargon.   

Solution: Picture Driven Design.  Up front.  Yeah!  I’m all for up front design!  A picture tells a thousand words.  It has the power to remove ambiguities, clarify the vision, showing what the journey to realise outcomes look like.  Start with a day in the life of…  scribble out the flow, the user journey.  Nothing complicated, boxes and arrows.  Then scribble out what the user interface might look like.  Done.  That’s your up front design.  That’s your documentation.  That’s your scope.  What you do next is up to you, write loads of documents that describe it and produce the software in a waterfall way if you want.  I’d prefer you were more lean and adopted agile practices, but whatever you do, start with the picture.  I’m convinced it will save much pain later on.

Marcelo Calbucci (Via Luke Wroblewski) describes a barrier that everyone assumes just has to be there so it is not tested: The CAPTCHA.    When the team removed the CAPTCHA from the Sampa.com registration form they recorded a 10% increase in conversion rates.  Marcelo writes “we created a set of tests and rules that will make us not display a CAPTCHA about 99% of the time” (interestingly when I looked at the page I got the CAPTCHA).  A 10% increase in conversion is not insignificant, and probably more than outweighs the additional effort to implement an alternative solution.

Can I email it to you later you asked her?

Silly question.

Have you ever tried emailing someone in the branch network?  Chances are the frontline staff don’t have external email.  What sort of business is it that doesn’t give its staff email and the ability to communicate on a personal level with its customers?  the same sort of business that has centralised all its customer contact and put it behind IVR.

So you give up on email.

Can I give you a ring with the details later you asked her?

She shakes her head.  She doesn’t have a number.  you see the branch doesn’t give out its number to the public. Go through the 0845 central switchboard, and let someone in the call centre pass a message to the person in the branch.  (With no formal process to get this to happen, good luck if your messages ever gets passed on).

In fact the only way you will probably get a message to the  person you talked to this morning is if you send a fax.  One step removed from the telex, two away from the telegram.

What is it with the banks that they don’t want their staff to have external email access?

Where is the value in that?  Where is the value in modelling a solution that is based upon that the current way of working?  What can you really learn about the business intent from the current as-is process?  A process that is based upon technical decisions, practical constraints and inherent complexity that were made at the time the software was built (in the enterprise world that was an average of 7.2 years ago), and hasn’t technology moved on since then?  By specifiying a system in terms of the curent reality the BA is missing a trick, and in the process is failing to add the value that the developement opportunity offers.  Indeed the BA is probably destroying value with over complex specifications that do not get down to the heart of the business need.

If Business analysts really want to create value, rather than dwelling in the past they should start with the “to-be” vision.  What are the desired outcomes of the application?  What should it do? (Not how should it do it).  Challenge existing assumptions, ignore perceived limitations and start with ‘blue-sky’ thinking.  What would we really like the application to do?  What are the customer / user goals and how can the application most eficiently (and delightfully) achieve these.  Think about the what, not the how.  Unlike the usual pattern of analysis, spend most of the time in the ‘to be’ world, only visiting the as-is to ask questions and confirm assumptions.

But that is only the first step of the BA creating value rather than destroying it. Why not go further and question the fundamental business model.  Rather than replace an old system with a new one, is it possible to reinvent the business itself?

Let’s take an example.  The client’s core business was data.  They supplied that data to their customers through an old and cumbersome desktop application.  They employed a service team to visit customers and install the system.  Data updates were sent via CD-ROM.  Their initial requirement was to maintain the desktop application.  The steer was to rebuild its functionality in an updated architecture (only one person in the company really understood how it worked, so updates were infrequent and key-man depecncy was a real concern to the business) develop a new local database and provide the ability to download batch data updates via the web.

The business analysis could easily capture requirements based upon this vision, yet they would be doing the client a diservice.  It would be possible to design and build a new desk-top aplication that woud be installed by the service team.  The technology would be better, database querries would be faster and customer attrtion would be slowed.  But it would be missing a trick.  The desktop application that IT wanted to rebuild was not where the value was.  The value was in the data.

Understanding what clients actually wanted and how they used the data (they used the desktop application to mine the data then copied and pasted into excel) it was possible to begin to challenge the existing assumptions.  The desktop app had charting and graphing functionality - why rebuild that when Microsoft do rather a good job of that with a product called Excel.  The requirements slowly changed into the ability for clients to pull data into excel via a web service.  It now would be possible to charge for data usage enabling creative new packages to be offered to different customers.  Previously customers had requested for changes to calculations and models to be implemented on updates for the desktop application.  Now the business started thinking about building a community where customers could create and share excel models and calculations on community website.  IT was helping the business create new value, setting aside the current reality and entering a domain of new possibilities.

The business analyst moved on from being a requirements taker to a change agent.  This was possible only with the mind-shift away from paying lip-service to the way the current application worked and thinking in terms of what and what if.