Selling agile is hard, if not impossible.

March 10th, 2007

A prospect call me yesterday. He wanted us to build a system that offered stock quotes, performance, history and whatnot. An ocean of information to investor. So far so good. But he had a deadline “it has to be completed by September”. I said that’s not how I work. He would be able to buy our hours, the system would be build incrementally. Every Friday or Monday he would see what’s been done during the week. He would be able to correct, change priorities, explain more about features. If the time is running out, you could select the main features. “No I want the whole system by September”. Serenity now!

“I had two quotes from companies that guaranteed the system would be done by September”. Great, they can guarantee that? I can guarantee that I will deliver an Island in central pacific with your system, in September, I said. “I can tell you anything in order to have you signing the contract, but that’s not how I work. I just cannot promise that the system can be done by September, we cannot anticipate things in software development, and ought to be surprises in the way”, I said.

And there would be. The system should have an online channel with Bovespa and BM&F. He wanted the system to update the quotes directly on the database, no lag. Imagine thousands of changes per second on several tables, performances being updated and all of this with three thousand users sending their queries. I’m sure we would have to be very creative to prevent timeouts and deadlocks.

I also mentioned that all my team were allocated this month, some of them in more than one project. So I wasn’t sure if I could start the project soon. “You can start the project in August, as long as you deliver it in September”. Ok, my patience doesn’t last forever.

I couldn’t lie to him or promise something that I just don’t know for sure (if the system could in fact be done by September) and he decided to go with one of the others companies, which btw charge more than us. This was a prospect from Brazil, and (un)fortunately I don’t get much of them. I had no problems convincing people from abroad about our approach, but most of them come from a technical camp.

I wonder if selling agile-built-projects is possible at all. At least here in Brazil.

Categories: Business, XP | Top Of Page | 19 Comments » |

19 Responses to “Selling agile is hard, if not impossible.”

Luke Melia Says:

You did the right thing, hammett. He didn’t understand now, but perhaps when he runs into problems on his project, he’ll start to wonder if maybe that Stronghold guy was onto something.

On the XP list recently, someone asked, “What’s the matter with me giving developers a date and a goal and asking them how they’re going to meet it?” My favorite reply: “What’s the matter with me saying, “You have to climb Mt. Everest in 3 hours, tell me how you’re going to do it.”?

hammett Says:

Hey Luke, thanks. I followed that thread, and was impressed with the question and the reply too. See http://hammett.castleproject.org/?p=127 ;-)

Mike Says:

I agree with Luke. You did right.
This is interesting. It never occurred to me that integrity is somehow bound up in agile programming mindset, but this seems to point to that.
It’s not uncommon in business for project managers to be able to estimate deliverables, though, so how do we reconcile the iterative methodology with (reasonable) expectations by business owners for delivery dates?

hammett Says:

Yes, estimation is a problem. I never been good at it, and always paid a big price for it (usually working from 10-12h/7 days per week). I always wanted to buy “Software Estimation: Demystifying the Black Art”

But back to your point, XP proposes a process to estimation. And I confess I never used it.. Need to re-read ‘Extreme Programming Installed’

Jason Says:

There will always be business people who don’t understand technology. It would be interesting to see if his project is completed in September. From what you have described, it would definitely be a challenge for even a large group of developers to complete in that time frame. And, if you HAD promised him everything he wanted, he would have most likely been on your back every week(perhaps daily), instead of playing the role of the customer and giving feedback which would assist moving the project forward. You did right.

hammett Says:

Thanks, Jason. I was thinking last night how much this positioning can affect my business to some extend. I mean, I cannot reject all prospect’s projects just because I cannot convince them. My impression is that people see software construction as something that will, for sure

- be late
- not well done

Maybe the challenge is showing how agile practices solve the second bullet, and the first bullet is up to the customer and will depend on the circumstances.

*sigh*

Bill Pierce Says:

If it were me I would call this customer on September 1 and ask him how well the system delievered by the other company is meeting his expectations.

hammett Says:

You’re funny, Bill. :-) But I think it would be a waste of energy.

Carlos Says:

That happened to me all the time in the Canaries, hammett. Some people is not ready for IT and is not gonna be ready for a long time, maybe never.
IMHO the best choice is to try to find out your place in other markets thru the internet as I think you are doing.
To move to another city could be also a good idea.
Best regards from Dublin ;)

hammett Says:

Hey Carlos, enjoying Dublin? I hope so. Moving is not an option, maybe I just need to hire some good sales people. :-D

MesBlog Says:

Praticare professionalmente l’agile: a volte impossibile

Alessandro Riolo Says:

I do believe Bill is right, Hammett. You must call this customer, if not in September the 1st of October, and you have to ask him how is going its project. If this will probably not gain you this work, it may help you to gain some future project with that same customer (then, I don’t know if that’s worthy the effort, it depends from the customer, of course).

hammett Says:

Hey Alessandro. I’m not sure that’s the kind of client I want in my portfolio ;-)

At least, if I can choose…

Dan Bunea Says:

Hi,

Every customer I’ve had, for which I haven’t worked before, ran like hell when I told him about a time and materials kind of contract, where we deliver every week what he wants.

On the other hand, a fixed price/scope project as you mention, is very risky, so what I usually tell my customer: target cost project: By september, with 3 people in the team the cost would be: X. We listed all your requirements, we gave estimates for each, you prioritized them, then we made a plan which reflects the features to be delivered every week (I wrote about this at http://danbunea.blogspot.com/2006/04/agile-adaptive-planning-and-fast.html).

Did you reach the point where this iterative delivery plan was done? Usually this is what gets my customers to calm down and see things with better eyes.

Thanks,
Dan

Juliano Says:

Hi Hammett

Oh gosh! How many times are we going to talk about the same problems again and again and again?

People need to treat software as a thing no matter what you say. You can always blame the last company that failed to deliver the projects on schedule and try again to apply the same model to other team. And fail again!

Even in companies that yell to everybody else that they do XP, I’ve seen lots of promises, schedules, price for the ones that are on time and some sort of punishment for the ones that are not (the majority). Unfortunately, that’s the world I live.

But I’m sure I would’ve done the same thing you did if I were on yours shoes.

Btw, why moving is not an option? ;o)

Take care,
Juliano

TestDriven.NET by Jamie Cansdale Says:

Reflector on Mono 1.2.3

Miguel de Icaza writes: This time Reflector will work out of the box with Mono on Unix (no special handling

Casey Says:

As long as someone is willing to come along and tell a client what they want to hear, I think the world of development is pretty much chaos.

A long time ago I did some work for one of the largest UK companies, I developed a large proportion of their web site single handed, over a period of 6 months. It was and still is one of the top 20 or so commercial web sites in the UK. I did that as a contractor for them.

Then they changed ownership, and they wanted (of course) to redevelop their web site. Their current outsourced supplier for the same web site charges them over £30 million per year to maintain and do some ongoing development work.

They have a second approved supplier too.

About a year after I left them they asked if I could quote for a couple of projects they had coming up, one was a 15-20 page ‘brochure’ site and one was a web site to manage a world wide conference. Both had got quotes from their internal suppliers of ‘way in excess of 6 figures’.

I quoted 2 weeks for each of them. Easy sites to do. £8000 each.

I was told they would love to use me, but as I was quoting way too low compared to their approved suppliers, they couldn’t use me. I had to increase my prices.

After a bit of surprised discussion, I finally agreed to charge them for 6 weeks work on each site, because it was as low a figure as I could get them to agree to. £25,000 per site.

The first site took around 5 days work.

The second took 4 days work.

Moral of the story is: business people know nothing about development, so they will just make a decision on who tells them the closest thing to what they want to hear.

I still haven’t found a client I can convince (properly) that Agile is the best solution for their development processes, despite it being the lowest possible risk for them. They all make overtures about how some aspects of it sound good, but eventually they all seem to slip back into their old ways.

Eduardo Miranda Says:

Well you could do just like the big consulting companies: Play waterfall, get an early sign-off and then blame the requirement changes!

Now serious, I think there are two kind of customer that don’t accept any risk of software building:
1 – The one who doesn’t understand anything about the software building and he won’t be ready to a agile approach because of that
2 – The one who has been there, done that and didn’t like the outcome (over budget, delays and bad quality) and resents that. This one is worse, because he will treat like his enemy from day one and will try to squeeze every requirement he can from you. He knows you will delay and he is actually willing that, because he going to use it against you.

For me they bad customer to rely on (if can choose, of course). But the second is worse because we will dry all your profit.

New serious, this is a bad project to you. When a customer wants to outsource all the project inherent risk he is usually

Console.Write(this.Opinion) Says:

A vis

A vis

Leave a Reply