Room to Breathe
The BA Role in Project Commitments
Robert T. Merrill, Principal
uFunctional LLC
You’re about to learn how to…
• Facilitate estimates using
The Stretch
• Talk about targets, estimates, and
commitments, and
shape realistic expectations
• Prepare and use a requirementsbased project history for
parametric estimation
Signs of suffocation
•
•
•
•
•
•
•

Schedule slips
Cost overruns
Forced overtime
Half-finished features scrapped
Late-lifecycle tasks cut short
Poor quality
Net negative value
In the beginning was the ship
date, and the requirements
were without form, and void…
Inception 1:2
The Stretch
100 1. Visual Aid
90
75 2. Outcome
3. Team and Tools
50 4. Suggest crazy low number
5. Suggest crazy high number

25
10 6. Repeat until you have a curve
0
Estimate to achieve Outcome
Product Concept
GreenRide promotes carpooling.

A user admin creates employee
accounts.
Employees enter their address, work
start and end times, and vehicle info.
Employees see everyone on a map,
marked as being a carpool member or
not.
Product Concept
Employees can join a carpool or start
their own.
Employees mark days when they
carpooled, and see how much they
saved on fuel and CO2 emissions.
Product Concept
Team: Two strong
ASP.NET, HCJ (HTML/
CSS/ Javascript) and SQL
Server programmers.

%
100
50

0

0

MO

Prepare a Stretch chart
9 for yourself.

Then I’ll lead you through an
estimate.
The Stretch
100
90
75
50
25
10
0
The Stretch
100
90
75
50
25
10
0
The Stretch
100
90
75
50
25
10
0
The Stretch
100
90
75
50
25
10
0
The Stretch
100
90
75
50
25
10
0
The Stretch
100
90
75
50
25
10
0
10%

50%

90%
The Stretch
100
90
75

8 mo
4 mo

50
25
10
0

2 mo
10%

50%

90%
The Stretch
100
90
75

8 mo
4 mo

50
25
10
0

2 mo
10%

4 months 50%,
2-8, 80%
50%

90%
Expressing Estimates
For the proposed scope and team, the
project has a 90% chance of being
done as scheduled by the end of Q3
2011.
Given three months and the named 2person team, the project has a 75%
chance of delivering the defined
scope.
Expressing Estimates
As defined, the product cost has an
80% chance of being between $100K
and $400K.

If it doesn’t come with a
probability, it’s not an estimate.
It’s something else.
Four Brutal Facts
1. Ranges and probabilities will be
unacceptable to someone.
2. The size of the range will be
unacceptable to someone else.
3. The only way to reduce the range
is to reduce sources of
uncertainty.
Four Brutal Facts
4. Reducing sources of uncertainty
takes:
•

Decisions that stick

•

Time

•

Money
I just need a rough estimate. I
won’t hold you to it. —Pinocchio
So it begins. —Théoden
Key Definitions
Target: Desired outcome
I’d like the shopping feature in time for holiday peak.

Estimate: Likelihood of an outcome
The team has a 50% chance of finishing the shopping
feature by December 1.

Commitment: Agreement to an outcome
Boss: It’s live by November 1st, or else.
Team: OK, we’ll get it done somehow.
Desk-Pounder Scenario
Negotiated Scenario
A famous, misused chart
4.0x

(The Cone of Uncertainty, from Software Estimation)

1.0x
0.25x
A famous, misused chart
Total
Uncertainty
(The Cone of Uncertainty, from Software Estimation)

(Scope &
Effort)

4.0x
1.0x
0.25x
A famous, misused chart
4.0x

(The Cone of Uncertainty, from Software Estimation)

1.0x
0.25x
A famous, misused chart
4.0x

Total
Uncertainty
(The Cone of Uncertainty, from Software Estimation)

(Scope &
Effort)
Initial
Concept

1.0x
0.25x

UI Design
Complete
The Estimation Problem
The Estimation Problem
The Estimation Problem
The Estimation Problem
The Estimation Problem
The Estimation Problem
The Estimation Problem
Your Mission…
Crossing the Hot Zone
• Understand it yourself
• Encourage sound thinking
• Negotiate well
– Delay commitment
– Commit near the top of the Cone
Crossing the Hot Zone
• Improve
– Estimation (zero bias, real ranges)
– Process—reduce uncertainty sooner
– Project management
Sources of Uncertainty
Requirements
We don’t know what it needs to do.

Team
We don’t know how long that’s going
to take.

Technology

Technology know if that framework and
We don’t

those components are going to play
nice.
Uncertainty is Increasing
Requirements
Volatility

Team
More roles
Virtual teams

Technology
Technology
Languages

Components

Platforms
Frameworks
Crossing the Hot Zone
Crossing the Hot Zone
Crossing the Hot Zone
Mistaken beliefs
Stretch goals
Gambling and suboptimal results
Parkinson’s Law and Student Syndrome
<are linear; suffocation is exponential

Trade quality for speed
Haste makes waste
Just add people
Mythical-Man-Month (1960s)
Common phrases
You hear…
I’ll find
somebody
who will!

You might say…
That sounds like a
threat to me (Pause).
Especially if my job’s on
the line, I can’t afford to
risk my reputation by
agreeing to something
that I don’t think is true!
Common phrases
You hear…
But I’ve
already
told
the CEO.

You might say…
It’s good we’re finding
out so quickly that’s
probably not achievable.
Would you like some
help explaining to her
why, and looking for the
best option we do have?
Common phrases
You hear…
There are
business
realities
here.

You might say…
Reality is good. Help me
better understand the
value proposition here.
We can figure out
something that will
deliver value in that
time frame—more than
a failed project would.
Common phrases
You hear…
I’d like to
hear more
of a
‘Can-Do’
attitude
from you
folks.

You might say…
I’m real big on keeping
commitments, and I’ve
found it’s best if I don’t
make commitments I
can’t keep. I’d sure like
to help you set goals for
this project that we can
all commit to.
Common phrases

The mighty oak was once just
a little nut that stood its ground.
Do the math. —Anonymous
Parametric estimation
At its simplest:
Hours per
Hours= Countable X
Thing

Or fancier:

Number of
Countable
Things
Countable things
1.
2.
3.
4.
5.
6.

Related to effort
Available early
Low overhead
Countable from software itself
Not too ambiguous
Makes non-technical sense

I’ve used: Lines of Code (LOC),
Use Cases, Use Case Points, &
Function Points
Making a parametric estimate
1. Count (or estimate) the
Countable Thing (CTs)
2. Look at Hours per Countable
Thing (HCT) from comparable
projects (size, team, technology)
3. Multiply
4. Estimate range
90-100 Countables in Greenride

Example—Project History
90-100 Countables in Greenride

Example—Project History

Designer
s arguing
90-100 Countables in Greenride

Example—Project History

Designer
s arguing

20 Reused
90-100 Countables in Greenride

Example—Project History

Designer
s arguing

20 Reused

Most like “N”
Example—Reasoning and Math
HCT=12.5-20%=10.0

HCT =___ to ____
Example—Reasoning and Math
HCT=12.5-20%=10.0
Range 1.75x, so 5.7 to 17.5

HCT =___ to ____
Example—Reasoning and Math
HCT=12.5-20%=10.0
Range 1.75x, so 5.7 to 17.5

5.7
17.5
HCT =___ to ____
Example—Reasoning and Math
HCT=12.5-20%=10.0
Range 1.75x, so 5.7 to 17.5

5.7
17.5
HCT =___ to ____
1700 17.5
H 540 =___
5.7

x ____
95
Example—Reasoning and Math
HCT=12.5-20%=10.0
Range 1.75x, so 5.7 to 17.5

5.7
17.5
HCT =___ to ____

1700 17.5
H 540 =___ x ____
5.7
95
6
1700
540
150
M 2 =___ / (2 x ____)
Example—Reasoning and Math
HCT=12.5-20%=10.0
Range 1.75x, so 5.7 to 17.5

5.7
17.5
HCT =___ to ____

1700 17.5
H 540 =___ x ____
5.7
95
6
1700
540
150
M 2 =___ / (2 x ____)
80% 2-6 months

50% 3 months
Example—Reasoning and Math
HCT=12.5-20%=10.0
Range 1.75x, so 5.7 to 17.5

5.7
17.5
HCT =___ to ____

1700 17.5
H 540 =___ x ____
5.7
95
6
1700
540
150
M 2 =___ / (2 x ____)
80% 2-6 months

50% 3 months
Parametric estimate—summary
•
•
•
•

Independent check
Negotiates well
Requires project history (3 or more)
Project history is easy but no one
likes to maintain it

The best time to plant a tree is 30 years
ago. The second best time is today.
You now know how to…
• Facilitate estimates using
The Stretch
• Talk about targets, estimates, and
commitments, and
shape realistic expectations
• Prepare and use a requirementsbased project history for
parametric estimation
It’s tough to make predictions,
especially about the future.
—Yogi Berra

Room to Breathe: The BA's role in project estimation

Editor's Notes

  • #2 WelcomeEstimation is Technique 10 in BABOK but not Top of MindShould be—we’re uniquely placed to set projects up for success, or notToday I’ll give you some tools and techniques to give your projects Room to Breathe
  • #3 Stretch—anything, anytime, forces RANGE.Client told me years ago, you’re not bad estimators, you’re bad negotiators.Powerful technique that holds up well is parametric—people are afraid but shouldn’t be.
  • #4 Signs of suffocationProject that wasn’t worth doing, strained relationships, turnoverWilla CatherOld anti-smoking commercial—”It’s enough to make you SICK. Isn’t it enough to make you STOP?”
  • #5 (Pause)We’ve all been here.Dates and budgets form early.Business sponsors don’t like ambiguity, like to agree on the numbers, do the deal, and let their people work out the details.Budgets and schedules are simple; requirements not so much—arcane language.Three things—first and last about estimates, middle about using them—amount of uncertainty, causes, how to come out of the Hot Zone with room to breathe
  • #6 Stretch--Learned? Invented? Help others estimateVisual, Freeing, Express uncertaintyExplain, then doWhiteboard or notebook—describe Done—easiest to think duration, so need a team sizeStretched to low for NO, High for NO PROBLEM, now fill in the middle until we can draw a curve
  • #7 Describes project outcome
  • #8 Describes product concept (cont’d)
  • #9 Describe team and tools.
  • #10 What’s the chance they’ll be done in a month?
  • #11 Nine?
  • #12 Two?
  • #13 Five?
  • #14 Three?Curve.
  • #15 Multiple pairs of numbers and probabilities.
  • #16 50/50 about Four, 80% from Two to Eight (4x).
  • #17 Range embarrassing and asymmetrical. That’s real.People often talk about +/- months. That’s wrong. +/- percent is a little better, but not much. It should be times or divided by some factor, in this case 2x.There are a lot of things that can cause a project to take longer, but a hard limit on how short it can be.
  • #18 Scope, Team/Effort, Duration/Effort. Define two, express third as a number or range with a probability.Something about units. Deadline in quarters, not days. Budget in $100K, not to the penny.
  • #19 Scope, Team/Effort, Duration/Effort. Define two, express third as a number or range with a probability.Something about units. Deadline in quarters, not days. Budget in $100K, not to the penny.
  • #20 Range/% flip out Type A’s.Size of range flips almost everyone—they’ll ask for a smaller range.Reducing range=reducing uncertainty. No shortcuts
  • #21 You can’t just make the range smaller by making it so. You have to reduce uncertainty.
  • #23 If you only remember one thing from the “Shaping Realistic Expectations” this is it. You have or express targets or goals.You assess a given outcome to produce an estimate.You ask for, negotiate, and accept commitments.According to Merriam-Webster, “commitment” is also “a consignment to a penal or mental institution.” That’s what we’re trying to avoid.
  • #24 Two scenarios on how targets interact with estimates and turn into commitments.No range estimate, no good decision support given to boss (but they didn’t encourage it)Bad commitment-&gt;project suffocatesThe outcome is something NO ONE would have picked, nor is it anywhere near the best that could have been achieved with the time and people available.
  • #25 Here’s a better way.Validate (or not) the target with an estimate.There’s a misfit, so clarify importance of scope or date.May be a little short of breath, but still delivers value—a lot more than the forced, bad commitment.
  • #26 Most famous chart in software estimation lore.Tells you how big your ranges are, and how fast they shrink through the lifecycle.BUT, this isn’t you—this is PAR for the course, and there are a whole lot of assumptions built in. PAR is very, very good. A golfer who shoots PAR consistently—a “scratch” golfer, is VERY good. Chances are you aren’t a scratch estimator. You’re worse, or worse yet, you don’t even really know.
  • #27 Getting this good requires a solid estimation practice, AND ALSO solid project management &amp; governance. We’ll look at where that uncertainty—the width of the cone, in a little bit, but it doesn’t go down by itself. If your sponsors and teams are out of control, it will NEVER converge.
  • #28 The author of the book I pulled this from, on the left, explains all this. If I could only have one estimation book, this would be it. The one on the right is also very good.
  • #29 Despite all of these caveats, we can learn some things about setting and negotiating expectations from the Cone.Let’s zero in on the part of the Cone where us BAs live.
  • #30 We get the initial concept today, and estimate it at 1100 FP, $200K, and 1 year, with a huge (16x) range. Best of all possible worlds. We’re scratch estimators.
  • #31 We run our usual intake process—a requirements workshop, a bit of technical investigation, and some exploration of possible teams. A little more than a month from now, and $6K spent, we’ve cut the range to 4x, and 50/50 estimate holds steady.
  • #32 4x range—Type A’s flip out—so we keep doing our BA shtick and tighten up the requirements. That takes another six weeks and $11K, and we’ve cut the range in half again, to maybe 1.5x-2x, but diminishing returns are setting in.
  • #33 We’ve used almost ¼ of our schedule and 10% of our budget, and we’re still outside that 25% safety zone we’d like to give the PM.
  • #34 What would it take to get to 25%? According to the Cone—pro-tour-level estimating and project control, remember?—we’d have to design the UI, and that takes us out another two months—1/3 of the schedule and pushing 20% of the budget.So another brutal fact is that diminishing returns set in well before our range is tight enough to guarantee room to breathe.
  • #35 Once diminishing returns set in, we hit what I call “the hot zone”—range too big for both sponsors and PMs, and shrinking it is going to require a lot more time and money.The actual project could fall anywhere within the cone. There’s a lot of room above that PM safety zone where the project could likely land and not have room to breathe.
  • #36 In reality, the Hot Zone looks more like this. Remember that the cone we’ve been looking at is for scratch estimators, working in organizations with pro-tour-level project governance.Your cone of shame is probably wider. That means that your projects stay in the Hot Zone a lot longer, and it’s a lot more likely that someone is going to freeze scope, schedule, and budget by decree. If the point they pick is outside that green bar, things are going to get ugly.
  • #37 You’re not going to solve this by being a better estimator (although that helps). It’s about education and negotiation.Your mission, whether you like it or not, is to preserve your project’s room to breathe through the hot zone.
  • #38 These are things you can do right away.
  • #39 This takes time, but you can make great strides in a year. Part of the first one is having a project history—which will also let you make parametric estimates, which will help you negotiate.
  • #40 You have to know WHERE uncertainty comes from in order to explain it, and alter your methodology to drive it out sooner.WHAT are we making? That’s BA stuff.WHO’s making it. “Resources” aren’t interchangeable; productivity varies 3x-10x; WHO isn’t just programmers, also the BA, UI people and…drum roll..the sponsor. 2.5x for same programmer/language/app size.HOW are we making it—technologies, code re-use, etc. The more productive a technology is, the more uncertainty it introduces. Huh?
  • #41 Getting better through driving out uncertainty earlier. Must know where uncertainty is to drive it out.Three basic reasonsWHAT are we making? That’s BA stuff.WHO’s making it. “Resources” aren’t interchangeable; productivity varies 3x-10x; WHO isn’t just programmers, also the BA, UI people and…drum roll..the sponsor. 2.5x for same programmer/language/app size.HOW are we making it—technologies, code re-use, etc. The more productive a technology is, the more uncertainty it introduces. Huh?
  • #42 Here’s how these improvements play out against the Cone.You start with targets turning to commitments. Any actual project outcome in the red is in trouble, and there’s a big chance of that.
  • #43 By committing later and higher, you shrink the Hot Zone to the orange area. Fewer projects will suffocate.
  • #44 By getting better (estimating better, driving out uncertainty sooner, and tightening PM), you can commit earlier and lower relative to the 50/50 estimate and keep the Hot Zone smaller soon.
  • #45 These are valid in some other areas of management, but in software they get you in big trouble.
  • #46 Most of the time these conversations don’t turn into a cage match, but be prepared.Role-play these lines if it will help you.Remember that a suffocated project isn’t in the boss’ best interest either, unless your environment is so political that your boss actually stands to gain from a failure.
  • #47 Remember your BA facilitation skills. It’s OK to ask for time to “think about it” and give people a chance to cool off (though hard to live with the waiting).
  • #48 By not telling the boss that you believe that the proposed commitment will likely result in failure, you are withholding information.By agreeing to a commitment you are endorsing something that you have good reason to believe is not in the company’s best interest.
  • #49 Can-do attitude (you’re yellow)Business realities (you’re selfish, you don’t understand business)I’ve already told so-and-so (good cop, bad cop)What do you need… (I tried to be reasonable, but…)What if you… (challenge your competence)So-and-so says it can be done in… (challenge your competence)I’ll find somebody… (open threat)Wouldn’t you agree that it’s easier to stay motivated and work hard towards a goal you’re convinced is attainable?Help me understand the business goal better…there’s probably something within your target timeline/budget that will deliver valueThanks for letting me know; at least we’re finding out right away that may not work out; I’ll be happy to help you explain what we think can be done in that time frameIt isn’t about the time spent on the estimate, it’s about the time spent resolving uncertainties about the project. When will we have a firm commitment on the team composition/architecture/requested scope reductions?I’ll be glad to go back over this with the team; that might help us to go faster, but every new thing introduces yet more uncertainty so it might not change our confidence at this point all that much.We do not live in a respectful culture any more. There are many respectful people, and many respectable ones, but popular culture does not honor them.Showing respect is a choice we can make; it’s easier to respect some people than others, but ultimately it’s our choice. If you cultivate this attitude all the time then you won’t have to fake it and fail when you need it. And this target—maybe they could lose their job over it.Negotiation is scary for us, and the people we’re talking to about schedules and budgets often do it for a living. But the best cure for our fear, and the way to maximize the chances of a good outcome for everybody, is doing our homework, mentally, relationally, and emotionally. Some of that homework involves establishing a reputation of being trustworthy. You can start today. Some of it means analyzing the situation and the possible outcomes and the pros and cons of each and the likelihood of each being achievable, and being able to talk about how you and your team arrived at those likelihoods. Some of it means visualizing the various ways the meeting could go, and preparing your response to various forms of non-useful behavior. I’m not very good at it, but I’ve come a long way. Do a web search on “sales closing techniques” sometime, and learn to spot them.Courage is not the absence of fear, it is doing the right thing despite fear. Courage comes with knowing you’re prepared. Courage comes with visualizing the results of a good agreement and a bad one. Sometimes it’s easier to stick up for others than it is for ourselves—think of your teammates and that will give you courage. Courage is also found on the high ground—if you’ve elicited, clarified, and validated the other person’s interests, it’s easier to propose an outcome that fulfills your interests if you know it also fulfills theirs, and then stick with it if they try for more at your and your team’s expense. Being prepared will help you elaborate the consequences. But don’t bluff or otherwise get clever. You’re probably not that good at it.I don’t want to give you the idea that every discussion of project estimates and commitments is a cage fight. Most are very cordial, and most project managers and sponsors really are trying to do the right thing. But they are often strong personalities, and often outrank us, and that can cause us to undermine their ability to do the right thing by withholding information, merely because we imagine that they won’t like what we have to say and will think less of us for it. On the contrary—most executives respect someone who comes prepared to respectfully present realistic options, and stands up to some pressure. If we know that a proposed commitment has a significant probability of leading to forced overtime, burnout, turnover, and quality problems that compromise the value of the outcome to the business, and we *withhold* that response out of fear of what they might say if we respectfully say so, we’re depriving them of information they need to make good decisions. Shame on us.The “no thank you” line is easier to get out if you’ve prepared and rehearsed it. “We’ve compared this project to four similar ones, and the numbers show that in order to hit the target date you’ve proposed, our team would have to double its productivity. That’s not a chance I’m prepared to stake my reputation on, and I don’t think the outcome of trying would be good for our business and for you either.”Remember your facilitation skills. If things are getting crazy, suggest a time-out. “Seeing how strongly you feel about this target, I’d like some more time to get together with the team and see if there are approaches that we haven’t thought of that would give us a good chance of having something worthwhile by that date.”
  • #50 More often than not, people will respect you for standing firm, especially if you can back up your position.Which leads to the third topic—parametric estimation.
  • #51 We’re now into our third and final topic, Parametric Estimation.This sounds esoteric and complicated, and it can be, but like mathematics, it’s very powerful and you can do a lot with a little of it.It can be applied very early and it holds up well in negotiations.
  • #52 Fancy scares a lot of people off of simple, which is really too bad.Count and ComputeMore accurate earlyNegotiates well (scope vs. task)
  • #53 There are no really great Countable things.
  • #54 Stay within ranges of Countable Things, Team sizes, project sizes, and you’ll be OK.As with almost anything, extrapolation is waay dangerous.
  • #55 Here’s an example for our hypothetical Greenride project. The best analog is N, and we’ll adjust for re-use. Sometimes it’s best to average several. There’s still subjectivity in a parametric estimate.
  • #56 Here’s an example for our hypothetical Greenride project. The best analog is N, and we’ll adjust for re-use. Sometimes it’s best to average several. There’s still subjectivity in a parametric estimate.
  • #57 Here’s an example for our hypothetical Greenride project. The best analog is N, and we’ll adjust for re-use. Sometimes it’s best to average several. There’s still subjectivity in a parametric estimate.
  • #58 Here’s an example for our hypothetical Greenride project. The best analog is N, and we’ll adjust for re-use. Sometimes it’s best to average several. There’s still subjectivity in a parametric estimate.
  • #59 Here’s the math worked out. It gives a slightly more optimistic answer than The Stretch did.
  • #60 Here’s the math worked out. It gives a slightly more optimistic answer than The Stretch did.
  • #61 Here’s the math worked out. It gives a slightly more optimistic answer than The Stretch did.
  • #62 Here’s the math worked out. It gives a slightly more optimistic answer than The Stretch did.
  • #63 Here’s the math worked out. It gives a slightly more optimistic answer than The Stretch did.
  • #64 Here’s the math worked out. It gives a slightly more optimistic answer than The Stretch did.
  • #65 Here’s the math worked out. It gives a slightly more optimistic answer than The Stretch did.
  • #66 Independence gives confidence, confidence helps you stand your ground.Separation of scope and productivity encourages the right kind of conversation and negotiation.Project history doesn’t need to be fancy. But it’s easy to procrastinate, at least for me.
  • #67 You now have two probably new estimation techniques that work well in the early lifecycle—diminishing returns and the Hot Zone. Neither takes very long to prepare.You also have some concepts around which to educate, and some tools with which to negotiate.
  • #68 If you move towards the early-lifecycle estimation and commitment problem, you can really help your teams and company a lot.