Scope Check those Ambitions

Any producer worth their salt has some spreadsheet that can crunch the numbers and spit out a date. This week we'll look at what goes into scope checker spreadsheets and how to use them to arrive at a headcount and a release date. No worries if you never used one, attached to this link is a working example using time estimates. Here is one using SCRUM points. Given the enormous amount of crunching in this industry, it's likely many companies don't use a scope checker like they should. This is the sanity checking step that keeps your team from biting off more than they can chew. 90% of the crunching comes from mismanaging scope at the planning state. For the other 10% we'll follow up at the end with a look at those behaviors and problems that can blow up scope during production.

The basic calculation is pretty simple. For each discipline category (art, code, design), given estimates and headcount, you'll want to arrive at the total months to complete their work. This is Alpha, which means feature complete and all estimated work is done. Then tack on time for the polish, QA and Cert and you arrive at a Release Date.

    (Total Effort / Head Count) / Standard Work Month = Months to Alpha        

Seems pretty straight forward, but the devil of course is in the details. Download this template so you can follow along. Here is the SCRUM points version.

There are 3 main sheets HEADCOUNT, ESTIMATES, and CALENDAR.

Head Count

Head Count These are all the engineers, artists and designers you have and all those you might hope to hire. You'll have a column for each resource type for each task. Don't worry about specialists. You can balance that out later. Exclude overhead people like managers and QA. Though important, their work doesn't drive the dates. Also skip the Audio folks as their work gets squeezed in during production and often during the polish stage. Then factor in the New Hires. This is the number you will undoubtedly tweak to arrive at the target date. Don't assume you can hire all at once. Enter in your Hire Rate and Delay in months. New Hires can also represent the people rolling off another project onto yours.

Work Month Months don't come in standard sizes. There are 52 weeks in a year, but 2 of those are typically holidays and another 2 are the average vacation time, so going with 48 weeks, or a 4 week month is a good rough estimate. You might also put in time off for company events or conferences or one sick day a month. You can calculate it as you like, but given an 8 hour day this typically amounts to a standard work month of 160 hours.

Allocated The allocation accounts for the fact that most people cannot put in 8 focused hours a day for a variety of reasons. Keep it real. Typical is 80 to 90%. You may find this differs by discipline to account for support issues or meetings. Consider this slack per discipline.

Estimates

Gross Estimates When scope checking, you gauge efforts in increments of work weeks for gross estimating. Put in estimates for each applicable discipline. Paint with broad strokes. Full task breakdowns with finer numbers like hours or days will take a lot of time and delay the analysis. This is NOT a schedule. This is scope checking done often before the design is even fully fleshed out. If you must, use 0.5 or 0.25 weeks as your smallest numbers. A single day is just not worth the frustration. Use the Quantity field for repeatable work. This makes it easier to adjust scope. Use the Off? field to toggle off a task. This removes it from the final hours calculations for purposes of outsourcing or scope cutting.

Use Historical Data If you have access to historical data from past projects that was well-maintained and accurate, those are absolutely preferred over brand new estimates.

Don't be Aggressive This cannot be stated loudly enough. Trying to make a date work by squeezing the estimates down is unrealistic and will lead to failure and frustration.

Estimates from the Right People Don't let the business people and program managers feed you numbers. Get your estimates from the discipline leads.

Use SWAG Scientific Wild Ass Guesses are educated estimates for rough estimating that should include a range of time given the unknowns and state of the design. Go with the longer time frame. If it's too wide a range, discuss the reasons and make some assertions about the design expectations to narrow the range and affirm the larger number.

Calendar

Start Date Go with the first of a month. This makes it clearer to see effort by month instead of split up. That's how the bean-counters see it anyway.

Slack Slack makes up for the lack of dependency delays and poor estimating. Do a minimum of 10%. Most PM's will advocate for slack even in the waterfall schedules. The assumption is that people will stay busy most of the time so this margin accounts for when they don't.

Risk On a particularly innovative project? Worried about hiring? Thinking your estimates covered only 80% of the work? Then by all means, put in 20% to the Risk factor. This is tacked on to the Slack factor to come up with the Alpha date.

Alpha This is all the estimated planned work being done. There is an alpha date calculated for each discipline, but the furthest out is used for the project Alpha field. You will be looking at these dates A LOT. If you don't see a date, your estimates may have EXCEEDED the 2 year calendar. Rather than add time, this is when you will look to add New Hires on the Head Count sheet and toggle Off some features or adjust quantities in your Estimates.

DO NOT BE TEMPTED to adjust estimates down without real analysis on those numbers. You only hurt yourself and your team by distorting reality at this stage.

Polish, QA and Cert These are the fixed estimates post-alpha (in weeks) that are used to determine the Release Date. Typical is 4 weeks for each, but you will know your team's track record and put in the appropriate estimates.

Critical Path Stating the obvious here for emphasis, this view is perfect for spotting the critical path in terms of discipline. Try to bring the alpha dates close to each other with design finishing first and code and art soon after. Art is where you want the critical path, because that is the easiest to outsource. In the example below, the 500 Level assets were broken up into 200 in-house and 300 for outsourcing, assuming you have an Outsource Manager and contractor.

Scope Implosion

No plan survives first contact with the enemy. -Helmut Moltke

Self-inflicted wounds during production can hurt the most. Despite all the best plans and successful hurdles in pre-production, it's the unplanned changes you didn't account for that can make timelines implode.

Crappy Estimating You did the planning, but the estimates you had were complete crap. Everything is taking twice as long. It becomes a choice of redesign and recommitment of scope with stakeholders ??, being late ??, hiring more people ??????, crunching existing staff ??????????????, or if the estimating was really crappy, all of the above. ????????????????????.

Unique Conceit This is where ambitious designers and artists want to make every game level and asset an innovative star of the game. The problem is that uniqueness, if not pre-planned, can cost a lot of extra time from everyone, and you lose the economy of scale.

Wandering Vision Some teams just can't help re-imagining their game, influenced heavily by a change in leadership or the brightness & shine of the latest games they're playing.

Blue Skies Forever Ideation is indeed the most exciting time of development, but it really needs to be contained to pre-production. Yet some teams can't stop ideation. This is distinct from ear-marking time for TBD features, like setting some coding time aside for minor level-specific code or tool improvements that comes up in the course of development.

Iteration-itis The work is never done because it seems no one ever signs off on it and moves on to the next level or chapter. This disease kills roadmaps.

Bait & Switch The goals of the JIRA story change, like Lucy moving the ball for the placekicker Charlie Brown as he sprints toward the football. Lack of clear and consistent goals on stories can really waste a lot of effort.

It Ain't Working as Promised When the tools or internal solutions don't work as promised, this reality check hits you like a baseball bat. Now you need a new solution or workaround that can be costly.

Most of these issues are addressed by more effective leadership and processes that hold leads accountable and increase their awareness of the impact of their decision-making or lack thereof. The rest can be addressed with risk-mitigation like padding estimates, bumping up the slack and risk factors, having backup plans, front-loading the riskiest work, and having proof-of-concept milestones on tools and the asset pipeline.

Until Next Time

Given the holidays and the author's own workload (scope check me please!), we'll be taking a hiatus until January and moving to a monthly schedule.

Merry Christmas and Peace out!