Estimating IT Projects - a Quick How-to

Naturally any post like this is going to be replete with over-simplifications and generalities. There are professionals who dedicate their careers to building solid and accurate project timeline and cost estimates. This post cannot hold a candle to such professionals. The purpose of this post is not to inform those professionals but to get the non-PMP employee thinking about ways to conceptualize IT-related effort and how to find value-oriented projects for their organizations.

Working Ourselves out of a Job

Finding candidate projects for automation and IT effort is simple to do provided you ask the right person to do it. There are people that have a job and like doing a job. And then there's the person to whom you want to ask the question, "What about your job are you tired with, bored with, thinking there must be another way to accomplish this?" The person that can answer this question is solid gold in any organization - this person wants to work themselves out of a job.

Estimating Value

Hard Value

The rule of thumb in this category is that claiming hard value delivered in a project moves the needle on the budget. There will be a budget line item that will change. Generally you will want to be able to tell the potential executive sponsor that the proposed project will "pay for itself in X weeks" or will cut X% off a particular LOB budget.

Hard-value delivered typically results from things like:

1. Reduced material costs to generate output (getting a printer that prints on both sides of a piece of paper).
2. Reduced recurring fee or licenses (developing automation to aggregate data feeds into a dashboard rather than letting all concerned parties aggregate the data on their own).
3. Reduced overhead (finding cheaper suppliers for anything from kitchen supplies to IT Cloud server space).
4. Green Projects (everything from opting for branded coffee mugs replacing styrofoam cups to offering all customer correspondence via e-mailed PDF).

Soft Value

 Value is "soft" when it is assumed that a dollar benefit will not materialize without other action taking place. Soft value projects tend to lump into two categories: efficiency (for now) or strategy (for later).

When a project has a soft value deliverable of efficiency we typically equate that to FTEs (Full-time Equivalent or Full-time Employees). You would speak of a project delivering 1.5 FTEs back to the business. And yes, there is an obvious way that this kind of project can move from soft value to hard value but I've never personally seen that happen anywhere I've worked.

Estimating the Project

Data Sources - 3 days per data source

Catalog all data sources for the project. If the data source is going to be "user input" that should be noted here but estimated in the Screens/Forms bucket below. A data source is any feed of data that comes from some other location. Any file from a website, or report from another system or dataset from the datawarehouse would be counted and cataloged here.

Data Stores - 1 week per data store

How many data stores does the project require be built? This is going to be the total number of tables or files that the new project will need to build or establish. Typically your data sources and your Screens/Forms provide the inputs into these tables for your project.

Screens/Forms - 2 weeks per screen

For each dataset you are trying to capture you will need screens through which you can manage that data. Typically you will need one CRUD screen per table.

Reports - 1 week per report

Each project requires output of one form or another. This bucket should include all business reporting, metric/KPI generation and due diligence reporting to demonstrate the project is meeting its proposed value.

Coming up with the Timeline

Data Sources and Data Stores can generally run in parallel but should be completed before work on the screens and reports. So ((X Data Sources * 3 days)+(Y Data Stored * 5 days))/Number of resources. For each person-week of work estimate $10K of cost. Forms and Reports work much the same way but you'd be smart to double time alloted as there will be changes to accommodate. ((Z Screens * 14 days)+(A Reports * 5 days))/Number of resources. Estimate the same amount of $10K per person-week for cost of labor.

You should now have a timeline as you append Screens and Reports number of days to the Data Stores and Data Sources days. If that number of days was 30 then for Testing/UAT you double that to 60 days. Again adding $10K per person-week.

Hardware and Infrastructure

Add $10K per server needed. If you need network upgrades and/or extension those have to be costed out and added as the biggest risk to timely delivery.

End Result

The idea here is not to get to a definitive estimate but more to let you and I know how big of a number of dollars that our business case must tackle. This estimation is crude but it allows us to decide whether the project is the size of a bread truck or a bread box. That is ultimately where the young manager-in-training's conversations with their direct managers will start.

"Hey boss, I have an idea for a project: 4 weeks and $120K all-in. It will pay for itself in 20 weeks all in hard-value. How do I get this project funded?"