Develop a reliable estimate during an elevator ride? Yes you can.

“Elevator Speeches” are an important tool in management. By chance, you find yourself in an elevator with a senior decision maker for 30 seconds. Can you do something useful with that time? You should have an “elevator speech” ready, or be capable of generating one instantly, for several topics, such as:

• Who are you?
• What do you do?
• What are you working on?
• How’s it going?
• Anything you’d like me to know?
• Do you think we could problem-or-concept-you’ve-never-heard-of-before-this-moment? What would it take?

This post talks about an important “elevator response” skill – quickly estimating costs by using rules-of-thumb worked out in advance. It uses Information Technology investment as the working example, but the concepts apply equally well to other fields, although you may need to recalculate some of the rules-of-thumb used.

You can develop estimates that are accurate within a certain range, and that are better than guesses, in a few seconds if you remember some rules of thumb, and if you practice and refine your technique.

Detailed cost analysis and proposal-writing is important, and takes time, and there is definitely a place for it. But fast estimates are used to make high-level strategic decisions more often than you might realize. You should develop this skill.

Nightmare Scenario

If this hasn’t happened to you, it will.

You follow another person onto the elevator and, as the doors close, you realize the other person is your Deputy Minister (private sector: company President), on their way up to the board room on the top floor.

The deputy says hello and you introduce yourself. Then,

Deputy:

So, what do you do?

… Oh, you’re the leader of the IT development organization? Great. I’m on my way to a meeting right now, and it just occurred to me that it might come up that we need a system to do requirement-you’ve-never-heard-of-before. What would that cost?

You:

I can send a business analyst to get your requirements right away, and have an estimate for you by end of week.

Deputy:

No, I’m meeting the Minister now. What’s your best estimate, now?

What often happens next

A number of reactions to this high-pressure situation are rather common, including, “No response”, “Rote inflated response”, and “Poorly informed lowball response”. Here’s how those typically go:

No response

You can think of all kinds of things that you don’t know right now. You don’t want to be anything but exact and perfect. So, you respond, “Sorry, Deputy, there’s no way to give you an estimate without going through our requirements and analysis process; we have to make sure we don’t miss anything.”

The Deputy says, “OK, thanks anyway”. When the opportunity arises later in the day, they remain silent, and the opportunity is lost. The Deputy remembers that you were unable or unwilling to do a fast estimate.

Rote inflated response

You can think of all kinds of things that you don’t know right now. You’re terrified that you’ll be held to an estimate that is too low and not have the time or funds to succeed; but “no one was ever fired for having money left over”. So, you “take a conservative guess then double it”. Essentially, you’re going to ask for as much as you think you can get away with. You can return whatever is left over at year-end, or use it for something else.

Most small systems probably cost $150K so you say, “It might be $300,000, and $100,000 per year ongoing?  Maybe more.” The inflection at the end of the sentence implies the “question mark”, re-enforcing that you’re really not sure.

The Deputy responds, “Really? Wow. OK.” Again, they don’t take the idea forward – it’s more money than they’re willing to invest – and the opportunity is lost. Or, even worse, they propose the idea, and are not believed. Your group’s reputation for inflated estimates increases.

Poorly-informed Lowball response

You can think of all kinds of things that you don’t know right now, but you hate to pass up an opportunity to get new work. You’re terrified that if you don’t get awarded this project, it might go to someone else and you’ll lose control. You know they sell a software package that does something like this at Best Buy for about $350 so it must cost something like that to buy or build. You’d need a computer ($3000 at Best Buy) and a couple of days to install it.

So, you respond, “Maybe $5000?”. (Again the spoken question mark.)

And the deputy responds, “Fantastic, thanks! Assume we’re a ‘go’ on this.”

Half a year later, a note written by the Deputy on a returned briefing note says, “You need another $50K and to hire another person? You said $5K!”  By the way, it’s performance review season.

What your response could be

You think for a few seconds while the elevator floor indicator ticks by: 3… 4… 5. Then, taking a deep breath, you say,

This is only an estimate – about $65K one time – that’s $35K salary and $30k capital – then $20K expense ongoing starting in year 2 for as long as the system is in use. There’d also be a small cost – around $10K – to decommission it at end-of-life. I’d need to get back to you on what current project would be sidelined to do this with existing staff. We could avoid that by using a consultant instead, for about $30K more. Don’t forget that the users might require training, and there would be an opportunity cost to take them off their jobs for that.

I could explain where those estimates came from if you like, or could get back to you with a more detailed analysis in a couple of days.

Looking thoughtful, the Deputy responds, “Thanks, that’s enough detail for now. I’ll get back to you if I need to follow up.” The project gets approved, and the actual costs come in within about 15% of this estimate. You’ve built up some trust in your estimating skills.

Where did those numbers come from?

Those figures were not fiction and they were not a guess. They were an estimate based on sound and defendable reasoning.

• The labour estimate assumed this simple system would take one developer 3 months. The 3 months came from your experience with small systems. (Know how many months small, medium, and large systems usually take.) $500 per day for 60 days gave us the $30,000. (See below for where the $500/day figure came from.)
• The capital cost assumed that we need to run this system on a server costing $10K. (Know your standard server costs!). We’d need 3 – development, production, and testing.
• The ongoing expense assumed 1 day every 2 weeks from an employee (which is 10% of a $100K person, for $10K), and assumed 20% of the server cost annually for renewal, for an extra $6K. We rounded the $16K up to $20K. (20% of purchase cost is a standard rule-of-thumb for annual maintenance costs of IT equipment.)
• 10% of purchase cost is a good rule-of-thumb estimate for the decommissioning of a system at end-of-life.

Do the homework for your domain

You can do estimates like that in a few seconds if you already know the various rule-of-thumb estimation values involved. These might be different for your field – but you can work them out, and you should.

Labour Costs

For labour, know how long typical tasks take, and know roughly what your staff costs are.

In this example, we’re talking about technically-skilled professional staff (CS or ENG). CS salaries range from $66K to $124K – but it’s a bell-curve since most of your staff are in the middle of this range. (i.e. you may have a few CS-01s and a few CS-05s, but most of your staff are probably mid-range, like CS-03.) If we do some kind of weighted average, we get an average salary of $93K. Add the 25% “benefits tax” that our finance group usually includes in calculating salary budgets, and we get about $117K as the annual loaded labour rate for a CS.

Do some simple math and it turns out that there are about 112,000 minutes in a standard work-year. That’s really close to the 117,000 dollars above, so the loaded labour rate for your average technical person is $1.04 per minute, or “a dollar a minute” to a pretty good approximation. That’s about $60 per hour, and just under $500 per day. Double that if you are using contractors – you are paying extra for their profit margin, their administration costs, and for not making a lifelong employment commitment.

$1 per minute for professional staff is a useful rule-of-thumb estimator for many purposes.  Use $2 per minute for Executive staff.

As an added benefit, this is useful to help you keep an eye on how long some decisions take.  If 5 technical people spend an hour deliberating, that’s 5 x 60 = $300 in labour costs.  5 executives: $600. Was that $600 spent making a $200 decision?

If your projects make heavy use of staff in a very different salary range, repeat the calculations above and know your per-minute, hourly, and daily cost for average staff.

Equipment Costs

If you are an IT manager, it’s important you know your average server cost; and don’t forget to allow for more than one if you need a development environment or spares.

If you are virtualized, or using Cloud, know the average cost to add an application, instead of pricing servers. If you are cloud-based, your cost will be expense instead of capital, and there won’t be separate “ongoing” maintenance or decommissioning costs – those are included in the Cloud service fee and are, in fact, a big part of the reason for using the cloud.

If you are not in the IT field, know what other equipment costs are often associated with your work. Office space? Vehicles? Specialized equipment? You might even have no standard equipment costs.

Grow your estimating skills using feedback

Do an estimate like this at the start of every project, even if you are not being cornered in an elevator. Do it yourself, with little or no research, before your experts do detailed costing and planning and fill your head with distracting details.

Then, after every project, compare what actually happened to your estimate. Of course it was wrong – it would be surprising if it wasn’t. But was it spectacularly or dangerously wrong? If so, why? What did you forget or estimate incorrectly? Learn and improve.

Important strategic direction-setting decisions are made with very coarse estimates all the time. Learn to use this process to your advantage by feeding it with good data.