1 To the future and back again: Long-Term Forecasting
2 1. THE CHALLENGE
3 Are we ready for build-out?As Gilbert approaches build-out, Council and city manager have placed emphasis on long-term planning Conversation starts with review of long-term financial picture Result: Long-term financial forecast Staffing model
4 - Jim Collins, author of Good to GreatStart by confronting the brutal facts. One good-to-great CEO began by asking, “Why have we sucked for 100 years?” That's brutal—and it's precisely the type of disciplined question necessary to ignite a transformation. The management climate during a leap from good to great is like a searing scientific debate—with smart, tough-minded people examining hard facts and debating what those facts mean. The point isn’t to win the debate, but rather to come up with the best answers…. - Jim Collins, author of Good to Great
5 How will the decisions of today impact tomorrow?
6 Critical questions to ask:Does current level of infrastructure maintenance set us up for long-term success? Will we have staff necessary to complete critical tasks as service base grows? Personnel ≈ 70% of operating expenses; will we have resources to hire and maintain staff?
7 2. EARLY ROADBLOCKS
8 Original forecast: Five year plans for financials Staff projected based on FTE per 1,000 New projections: Wanted more sophisticated approach Hired consultant in 2013/2014
9 Ideal: A. B. C. D. E. Reality: A. Q. G. W. B.
10 Consultant – staffing modelExcel-based Quickly discovered fatal flaws - Population based - All positions projected same - Lack of operational knowledge - Graphs ≠ meaning Mary + Kelly at end from consulting experience Hired consultant in 2013 Model was flawed in a few areas; lack of knowledge of operations Assumed growth in perpetuity Less ability to change between years A lot of graphs, not a lot of meaning For instance, they told us which positions had the most incumbents Gilbert decision Take offline and rebuild Research + take what we liked and didn’t like Use the players you have on the field
11 “There’s no way to project that.”- Everybody
12 “There’s no way to project that.”- Everybody So, our manager cleared the path and helped the organization see the importance of and our commitment to both projects.
13 3. BUILDING A DYNAMIC MODEL
14 Long-Term Financial and Staffing Forecasts Process and FrameworkProject through 2030 Project on baseline services - Do not assume Council will elect to add/change services - Build in flexibility to change services if they so choose Revisit assumptions annually No black box calculations; met with department directors and division managers for input Devil is in the details, but utility is at 10,000-foot level
15 Long-Term Financial Forecast: Three key components1. Basic Assumptions 2. Revenues 3. Expenditures
16 Q: What key components may change over time?Assumptions What inflation index should we use? What are population growth assumptions? How will personnel-related costs change? Forecast task: Build a section for overarching assumptions that can be updated easily
17 Assumptions
18 Q:How will revenues likely change?Use existing formats Spend time on highest value Remember some revenues decrease over time Forecast task: Start forecasting using general trends Add detail for specific high-profile scenarios
19 Revenues
20 Q: What if current trends continue?Expenditures Keep it simple! Group items that trend the same Spend the most time and thought on personnel Forecast task: Decide groupings first Fill in data Look at trends
21 Expenditures
22 Staffing Forecast: Three key projections1. Ratio + anticipated growth 2. Position weights 3. Manual adjustments
23 Current ratios + anticipated growthQ: What is an appropriate ratio or input that drives FTE needs? Current ratios + anticipated growth What drives new FTE? Division-specific indicators Reasonable growth assumption 0-5% in five-year spans Forecast task: Look at position level Project based on these assumptions
24 Court Services Clerks:Ratio example: Court Court Cases: 30,000 Court Services Clerks: 10 Ratio: 1 to 3,000 Growth Assumption: 2% Division Metric Input Growth Assumption
25 Court Services Clerks:Ratio example: Court Court Cases: 30,000 Court Services Clerks: 10 Ratio: 1 to 3,000 Growth Assumption: 2% Forecast result: 1 new FTE 2022 1 new FTE 2027
26 Position weights Forecast task: Categories: 1. New service levelQ: Should an individual position be weighted to recognize extra demand? Position weights Categories: 1. New service level 2. Increased demand 3. Legal or risk 4. Infrastructure / CIP 5. Counterbalance / negative weight Capture what’s not in ratio and shifts between service lines Forecast task: Weight positions identified Reset every five years
27 Review weights for 5 areasWeight example: GIS Technician Review weights for 5 areas Increased Demand: Low Medium CIP: Other Weights: Stable Position Select weight Result
28 Review weights for 5 areasWeight example: GIS Technician Review weights for 5 areas Increased Demand: Low Medium CIP: Other Weights: Stable Forecast result: 1 new FTE 2020 Additional layer added to show growth beyond ratio; may also be used to delete over time
29 Manual adjustments Forecast task: What would we need if largeQ: Is there a policy-level decision that must be made before this service is added? Manual adjustments What would we need if large projects come online? Also helpful for items where staffing is known, e.g. fire station Should be the exception, and not the rule Build in flexibility for scenarios Forecast task: Look at position level Project based on these assumptions
30 Identify Need and YearsManual example New Regional Park Identify Need and Years Populate Include? Y/N Forecast result: 23 FTE spread out over five years Utility built in to turn on or off this
31 How do the models interact?They talk!
32 Time to go live Show impacts in model
33 4. LESSONS LEARNED
34 Lessons Learned Commitment from the top. If our city manager hadn’t committed to getting this done and made it a priority, it wouldn’t have been done. Patience. Practice it early. Practice it often. It’s an iterative process. Law of diminishing returns. Effort and value on each of the axis – you won’t ever make it “perfect.” In fact, some areas work better at a macro level. Keep it simple. Don’t overcomplicate it so it loses its functionality as a policy tool. DO make sure the data and layers are correct, but make sure the reports are at the level you need. Compass. Remember it’s a compass, not a GPS. It will point you in a general direction. Think of this tool like the Tin Man. It’s made in Excel. It might be better than any of us at math, but it doesn’t have a heart or a brain. You need a human to tell it what to do. Target practice. The further out you get, the less accurate your aim will be. OK to run different scenarios and see what sticks. “Because the model told me…” is not a stand-alone reason. Use it as a starting place; something to look into – then needs to be validated
35 5. TIPS FOR GETTING STARTING
36 How to get started Begin with the end in mindStart with how you’ll need to report out Outline a manageable project and timeline Build a project team that can see the forest AND the trees Build assumptions that can be reset or changed over time Keep one master that populates everything What is going to be most useful for your community? You can always take out – it’s harder to put layers back in once you’ve built it Maybe start with one department or fund. Recognize it’s a marathon, not a sprint You need the data to be right, but you also need to build in functionality that will be useful for decision-makers Make it flexible. One assumption forever is not that realistic. Don’t make one change that has to be updated 100 places.
37 Remember why you startedSet community up for long-term success, and better understand impact of decisions You might not be able to predict the Cubs winning the world series within a year like Back to the Future, but you will be better prepared for long-term conversations.
38 Thanks! Any questions? For more information contact:Kelly Pfost, Management and Budget Director Mary Vinzant, Assistant to the Town Manager