Sid Kemp is a business consultant and author of 10 books on project management and business success.
The Need for Project Time and Cost Estimation
In business, the purpose of most projects is to make money. That requires that the project either earn more or save more than it costs to do the project. A project costs money until it is done, and only begins to improve the company's bottom line after it is done. Also, each year, a business has only so much money it can spend on projects. So, even before a project begins, executives want answers to four questions:
- How much will the project cost?
- Are we sure we won't run over budget?
- How long will the project take, and when it will be done?
- Are we sure we will not run late?
Those are very reasonable questions. We ask the same things when we take our car in for repair or get our air conditioner fixed. Yet well over half of all projects—even well-managed ones—run late or go over budget. So answering these questions with estimates we can rely on and commit to is a big challenge.
In the project management world, we call executives that commit money and organizational effort towards a project the project sponsors, and their commitment is crucial to our success as project managers. If they don't have buy-in, stay committed all the way through, deliver necessary resources, and remove roadblocks, our project will fail. And their request for a time and cost estimate is, from their perspective, entirely reasonable.
It is also very difficult, sometimes impossible, to provide. By definition, a project is "a unique endeavor undertaken to create a unique product or service" (PMI Lexicon). If each project is unique, and it's never been done before, then how are we going to know how much it will cost to do it? How do we know how long it will take? How do we make an estimate of the cost of the unknown? And how can we be sure we are right? If we're not right, we're going to look bad when we run over budget or deliver late, or both.
These are real questions, and they have real answers. I have taught over 3,500 project managers how to plan projects, estimate costs and time, and deliver on time and within budget. I've written three different project management books with time and cost estimation techniques. And, from my clients and students, I've seen everything from excellent, down-to-the-dollar estimates that can win competitive bids, to management systems that make a good estimate impossible. But good estimates are possible. And we gain a lot of authority that makes our jobs easier when we, as project managers, estimate our projects well.
In this article, you will learn how we, as project managers, can create accurate, reliable estimates. You'll learn:
- what we need to have to create an estimate,
- what types of estimates are possible, and
- the basic procedures for preparing each type of estimate.
Detailed instructions for estimation techniques are too large to fit into this article. But you will know the name of the right estimation technique and how to use it. And you will have an example to follow. From there, you can probably do it yourself. If not, you can easily find detailed instructions in a project management textbook, or in the PMBOK, the Project Management Institute's Guide to the Project Management Body of Knowledge.
This article shows you how to make the best time and cost estimates you can, both early in the project, and later, when planning is nearly complete, as illustrated in Figure #1.
A Computer That Runs All the Time
One of my clients provided computer systems to government agencies. Before I worked with them, an agency asked for a computer for their email system that would "run all the time." The project manager ran with that and bought a $1 million computer that would run 24 hours a day, seven days a week, with dual automatic backup of everything. When he told the client, they were shocked. "But we're only open 16 hours a day. We don't need it to run all night. And we're certainly not going to pay for it!"
The company was stuck with a $1 million computer the client didn't want. And the real solution cost much less than half of what the project manager assumed would be needed.
The lesson: Don't make assumptions, and don't assume you understand what your client means when they use certain language. Check every detail in a clear, definable way.
Prerequisites: What You Must Have Before Creating a Time and Cost Estimate
Imagine a friend came to you and said, "How much will it cost for me to buy a new car?" There is no way to answer that question without asking, "What kind of car do you want?" And what if he replies, "Oh, I don't know anything about cars. Just get me a car that will work for me. How much will it cost?"
In that situation, we can't estimate the cost of the car. We have to work with our friend, asking questions like "How big is your family?" "Do you drive to work?" and "What do you like to do on vacation?" If our friend works with us through all this, and we know cars, we can probably pick a good car for them, and then tell them the cost. But if our friend says, "I don't have time to answer all your questions! Just tell me how much a car costs!" then we're right back to nowhere. Or, we say something between $12,500 if you want a Hyundai and $85,000 if you want a Cadillac. That's a true answer, but not a very useful one.
As project managers, we find ourselves in this situation all the time. For example, a client may want a new website. But they don't really know what a good website for their business is. They just know that the current one isn't working. We know that a website can cost between $1,200 and $15,000, and take between two weeks and four months, depending on what features are needed for a particular business. But, until we know their business, we can't make a time or cost estimate.
The bottom line is this: Before we can prepare a precise time or cost estimate, we must work with the client to define the scope of the project. In case the term "scope" is unfamiliar, it means "what we are making." A full scope definition is a clear statement of what we are making by doing this project. To learn more about this see, How to Lay Out a Project Management Plan.
In addition to a clear scope statement, we need three more things:
- We need to eliminate unspoken assumptions. For example, our friend asked for a "new" car. Did they mean brand new, this year's model, or did they mean "new for me," and buying a used car would be just fine? We don't know until we ask. See the sidebar about "A computer that runs all the time" for a real case of a project cost estimate disaster like this one.
- We need to decide how the project will be done, and get an agreement on the basic approach from the client. We don't need to know all the details. But we do need to know the general approach. For example, if our project is to set up an assembly facility, we will want to know if the customer expects to outgrow it soon. If so, then renting property, rather than buying it or getting into a long-term lease, is probably the best idea. And the cost of a project involving rental property is totally different from the cost of a project involving buying or leasing a building.
- We need to know the current situation. If the client wants a new website, then we need to know all about their old website, and what works, and what doesn't. We also need to know about their printed marketing materials, product descriptions, and images of their products. Creating a website from a current catalog costs a lot less than creating a website from scratch, where we have to arrange to get photos and descriptions for dozens of projects.
There is a story that illustrates this last point. One time, there was a project to drain a swamp. A county executive asked, "how much will it cost to drain this swamp?" The project manager answered, "it depends on how many alligators are in there." "Well, how do we know how many alligators are in there?" "We drain the swamp, and see how many run out."
This is the big problem with time and cost estimation. We have to plan enough to know what the customer really wants, and do enough work on the project to discover what is really going on before we can make an accurate cost estimate. But, often, the executives want an accurate cost estimate before the project is even started.
Now that we've laid out the problem, let's take a look at some solutions.
Our Case Study: A Simple Website Cost Estimate
The amount of work it takes to make a good cost estimate depends on the size of the project you are doing, and also on how familiar you and the team are with project work. The estimation techniques described here can give you a good project cost estimate in a couple of hours for a project costing under $10,000. The same techniques, applied in a couple of weeks, can estimate projects costing around $100,000. And the same techniques, applied carefully with a lot of research, can be used to estimate projects that will cost millions of dollars. But that will take months of work, and a lot of other useful project planning will go into the cost and time estimate.
Because we are focusing on the techniques, we will work on a very simple, small project: The first website for a new startup company, including 40 pages of text and graphics, and a shopping cart with about 200 products. A print catalog with all product images and descriptions has already been done. But the company does not like the catalog layout and needs a new artwork and concept design for the website, which it will also use for future print marketing.
Table #1: Early Estimation Totals
Hosting (first year)
Early Cost and Time Estimation Techniques
As we have already said, the earliest cost estimate must come after the project scope is clearly defined, an approach to the project is agreed on, and assumptions are clarified. When this is done, but we do not yet have a step-by-step plan for doing all the work, we can create an early cost estimate. An early cost estimate is called an order of magnitude cost estimate by the Project Management Institute. When we create it, we provide a range in which we are almost certain that the cost of the project will fall. But it is a pretty big range. The low end is 50% of our estimate, and the high end is twice our estimate. So, if we think the project will cost $100,000, we can say with confidence that the project will cost between $50,000 and $100,000.
Presenting Order of Magnitude Estimates
When we deliver an estimate with such a big range, the project sponsors will probably go ballistic. Most project sponsors want, and need, a much more precise estimate much earlier than is possible. It takes educating upward, that is, teaching our bosses what is realistic, to help them understand that an order of magnitude estimate is the best that even the best project managers in the world can do, early in a project, before the plan is nearly complete. At this stage, there are simply too many unknowns to make a more realistic guarantee.
There are several pieces of good news for us and our sponsors about early project estimation:
- If we are doing a type of project that is familiar to the team, then we can be more confident that costs will not be too much higher than the center point of our estimate.
- If we can control scope, we can limit the project to the available budget by delivering some, but not all features.
For example, say we have created an estimate between $50,000 and $200,000, and the sponsor says, "I really can budget only $120,000, and, if it runs over that, maybe I can squeeze out a total of $150,000."
We can reply, "$100,000 is likely, and there's a very good chance that we will come in under $120,000. If you want a guarantee on the $150,000, then agree that, if we absolutely have to, we can cut two of the five items on the list, and we will do the project for $150,000 or less. You will probably get everything you want. If not, you'll get the three or four most important things, and we can do the last one or two on next year's budget."
Breaking our sample project into a few big pieces
Most projects can be broken into a few big pieces, and estimates put together from pieces are always better than estimates created from the project as a whole. So let's look at the pieces of the website project in our example. We will include:
- Hosting: Set up hosting and the wireframe for an eCommerce site with a shopping cart
- Graphics: Work up a new graphics design: colors, site image, logo, and key images for each of seven product lines
- Web Pages: Create 40 pages of text and graphics
- Catalog of Items: Create a shopping cart of 200 catalog items from existing product names, product descriptions, and images
- Go Live: Test the entire site, fix issues, and go live
Getting High and Low Estimates for Each Piece
Our first step is to figure out the best way to get a cost estimate for each of the five parts of the project listed above. For items #1, #2, and #5, our best solution is to create an estimate based on the actual cost of doing the same things for similar sites for past customers. For item #1, we can probably even lock down a fixed price. On item #2, we talk to the graphic artist. They may say, "Well, it depends. If they know the general style of the business, and they like my first ideas, that's one thing. But if they have no idea what they want, and they have a committee making decisions, and then they start changing the names of their product lines while I'm doing the graphics, it could cost a lot more." If we have good communication with the client, we can try to get those questions answered. But let's say we don't. Then we ask the graphic artists to estimate the cost for the best case and worst case scenarios, and those are our low and high estimates.
Items #3 and #4 are best estimated using the Bucket Method. But the method described for items #1, #2, and #5 will also work.
Item #5 is similar to item #2. For example, if all sales are in the Continental US, then testing a shopping cart is standard. If international shipping is needed, along with international taxes and tariffs, a lot more testing will be required. In this case, we went to the client and they said, "Oh, we're in the US and only selling in the US right now." So we got a pretty narrow range of low and high in our estimate for going live, as you see in Table #1.
This method puts together a good early time and cost estimation. The cost estimate is illustrated in Table #1. The time estimate is created in a similar fashion, asking each worker or team leader for a low, likely, and high estimate of how long the work will take, and the estimates are added together.
A key advantage of this approach is that, even if one part is estimated too low, another part is likely to be estimated high. So the total estimate has a greater chance of being right than any one part. This is the major reason for breaking a project into parts and estimating each one separately, instead of trying to estimate the entire project all at once.
Accuracy Varies With Industry
How accurate is our estimate? That varies a lot, by industry, and also by other project and team factors. In the construction industry, with companies assembling competitive bids to build roads for projects funded by the state, which generally chooses the lowest cost bid, I've seen a $25 million project estimated down to the last dollar. A bid that is too high won't get the contract, and a bid that is too low will mean financial loss for the company.
Construction projects can be estimated so precisely because methods, materials, and costs are very standard.
On the other hand, information technology (IT) projects are notoriously badly estimated. The worst case I know of was an IT project estimated at $2,000 that ended up costing over $170,000, but that was an extreme case. But actual results ending up double or triple the estimate for time or cost is far too common. One reason for this is that tools, techniques, and environments change rapidly in the IT world. It even happens that a program is written, say, for one version of Windows, then has to be rebuilt for another version of Windows before it is released.
How do we make the best estimates we can in our own industry? Consistent, complete project management planning and tracking and lessons learned to improve our methods at the end of each project are key. And, when doing each estimate, ask, "How familiar is my team with this exact type of work? Have we ever worked together before? What is new and different on this project?" The more often your particular team has done this type of work before, the more accurate your estimate can be. And, at the same time, when you are estimating work using new tools and methods, developing experimental prototypes, or working with new clients, take that into consideration, and allow more leeway for the unexpected.
Late Cost Estimation Techniques
Later in the project planning process, we can create a much more accurate estimate. And it is a lot easier to do, as well. The PMI calls late time and cost estimation bottom-up estimation, because we estimate each detailed activity, and add them all up to get a total. The PMI says that the actual cost will be -10% to +15% of the estimated cost using bottom-up estimation. And schedule variance is the same. That means that, if the estimate is $100,000, the actual project cost is expected to fall between $90,000 and $115,000. If a project is expected to run for 100 days from beginning to end, then it is very likely to run between 90 and 115 days. That smaller range is much easier for the project sponsors to work with.
Late time & cost estimation requires a detailed project plan including:
- A work breakdown structure (WBS) - a detailed list of all project components
- An activity list - a list of all the work that will be done, and who will do it.
- Detailed cost information, such as hourly rates for salary, contract work, and other items such as office space or equipment.
- A procurement plan, including estimated, or when possible, confirmed costs of every item to be purchased.
- A risk plan
Ideally, the detailed cost estimate will be part of a complete project plan, in which all nine areas of project management knowledge are addressed, as described in the article I listed above, How to Layout a Project Plan.
With all this information available, creating a detailed cost estimate is quite easy—in fact, it is easier than creating an early estimate. We simply add up the cost (time and materials) for each activity of the project and come up with a total. As part of this process, we can use a PERT estimate. PERT stands for Project Evaluation and Review Technique, and it is a way of getting a single cost estimate using a centering average of optimistic, likely, and pessimistic estimates for each line item. You can learn PERT by reading How to Use a PERT Chart for Project Management.
A best practice for cost estimation includes factoring risk into the equation. Part of risk management is to include contingency plans, that is, plans for work that we may do if a risk event occurs. Let's say that we're building a new building in Florida. We're aware of the risk of having to shut down the construction site for up to two weeks in the case of a hurricane. We plan a schedule that gets the building done, even if a hurricane comes through, by using weekend overtime after the hurricane. This, of course, adds to the cost. We can present risk factors as contingency funds. For example, we can say, "The estimated cost of constructing this building is $1 million. However, if a major hurricane occurs, we will need an extra $50,000 of contingency funds to finish the project on time."
Tracking Project Work and Adjusting Your Estimate
When your project plan, schedule, and cost estimate are accepted, then the planning work of your project is complete, and you can begin doing the work. Formally, this is called project execution and control. Execution means doing the work. Control means keeping track of the work, and making sure it is getting on time and within budget. The process of tracking work completed (scope deliverables) against time and cost is called Earned Value Analysis (EVA). As we do EVA, we work to ensure we deliver within budget. We also learn if we did a good job making our estimate in the first place.
Tracking actual time and cost as we work is very important for two reasons. First of all, that data is useful input for future estimates. Secondly, at the end of the project, we compare our estimated costs to the actual cost, line by line, and our planned schedule to our actual schedule, day by day; and create a lessons learned document. The lessons learned include better methods of cost estimation for our next project.
This article is accurate and true to the best of the author’s knowledge. Content is for informational or entertainment purposes only and does not substitute for personal counsel or professional advice in business, financial, legal, or technical matters.
PMP Geek on October 04, 2016:
A well written article. My question is how do we perform estimation for 'in-house projects' where the team comprises of paid employees. There is no additional cost involved for the project. The team members are assigned to a project and they need to finish it. Also, as a project manager, I may not have access to their salary details (HR looks after it). Thanks in advance.
Sid Kemp (author) from Boca Raton, Florida (near Miami and Palm Beach) on April 18, 2013:
Thanks, SoCurious. Your comments are both insightful. First, there are real situations where the project is "fix this broken system," and we can't really perform the estimate until after the diagnosis. Auto mechanics and brain surgery are two fields of endeavor where this problem appears frequently. This was *not* a problem during the early years of the Space Program when the engineers fixing problems were the ones who built the system. It *was* a problem in the final years of Space Shuttle maintenance when the systems did not have clear "as built" engineering diagrams and an indicator light said there was a problem - was it the indicator, the wiring, a loose connection hundreds of feet inside the shuttle, or a real problem with the hydrogen tank. That one delayed at least two launches by several weeks.
And as for people personalities - yes, these will *always* be a moving target. My solution is to teach each project team member to be self-led and self managed. You might check out my hubs on self-leadership and leadership for more.
SoCurious from Northern Virginia USA on April 18, 2013:
Well written article. Mechanics and brain surgeons often have to resort to using a "we won't know until we drain the swamp" approach to their customers/patients. There are so many variables to estimating and running successful projects, with people personalities being a moving target.