What's Amazon Operational Plans (OP1 and OP2)

Operational Plan (OP) the company's process to develop a long term business plan, which culminates in a business and financial plan for the following year. OP documents are classified in 2 parts: "OP1" and "OP2".

OP1 starts in Q2 and is focused on long-range planning for the business. OP1 documents are forward looking, they may span (depending on the org) three to five years and outline the future path for the business. OP1 sparks strategic conversations, leading to the objectives and actions necessary to achieve the goals spelled out in the document.
OP2 begins immediately after OP1 and is generally approved in Q1 of the following year. OP2 represents the official target for the business team. OP2 documents are focused on the immediate next year and provide details on trends, key metrics and business drivers.

OP1s and OP2s are used when:

Laying out the team's strategy for the next year and request headcount/ budget to support it.

Why do we use OP1s and OP2s?

As a standard process for operational planning at Amazon.
To think big about continuing vs. new projects and priorities for the following year.
To get into the details of headcount and budgeting needs; get approval from senior leadership.

Content and Structure of an OP1 Document

The below is a recommended high-level structure for an OP1 doc.

• Purpose of team/vision
• Financial summary
• Situation (performance, highlights, low-lights)
• Plans
• Goals
• FAQ:

1. What are the hotly debated topics in your area or most difficult judgment calls?
2. What “dogs not barking” worry you?
3. What are your disruptive ideas?
4. What were your positive surprises over the last year? How are you doubling down on them, and what are the specific actions and programs that evolved from these positive surprises?
5. What were your top misses and learnings? What are you most disappointed about in your area (dirty laundry)?
6. What are your top “paper cuts” (where a paper cut is a small issue on its own, but taken in sum with other similar-sized issues, diminish the customer experience)? What are your plans to address?
7. What hard choices were you forced to make to fit within the headcount targets? What items were the most painful for you not to fund?
8. What mechanisms do you have to reduce your fixed and variable costs as a percentage of revenue?

• Tenets
• Appendices (with data tables)

Example Structures

The following were the structures of real OP1s from the last few years:

Amazon DSP 2022 OP1

• Team Mission and Background
• 2021 Performance (highlights and how you are capitalizing, lowlights and how you are addressing)
• 2022 Plans (including 2021 budget, BIS, incremental HC ask)
• 2022 Goals
• FAQ

1. What are the hotly debated topics in your area or most difficult judgment calls?
2. What “dogs not barking” worry you?
3. What are your disruptive ideas?
4. What are your top “paper cuts” (where a paper cut is a small issue on its own, but taken in sum with other similar-sized issues, diminish the customer experience)? What are your plans to address?
5. What hard choices were you forced to make to fit within the headcount targets? What items were the most painful for you not to fund?
6. If forced to keep headcount flat year-over-year, what harder choices would you make?
7. What mechanisms do you have to reduce your fixed and variable costs?

• Team Tenets and Tension
• Appendices (with data tables)

Corp Apps 2011

• Mission/Vision/Background
• Tenets/Guiding Principles
• Current Performance
• Plans, Goals, Strategic Opportunities, and/or Big Ideas
• Categories of Investment:

1. Keeping the lights on – investments to keep the current businesses growing at their expected rate.
2. Efficiency of keeping the lights on – areas we can invest in to get higher efficiencies/lower costs in the things we do in the above area.
3. New business opportunities – investments to create new businesses we will own.
4. Support of new business opportunities from around the company – investments we are making to not just keep the lights on but to support other new businesses/growth opportunities for the company.
5. Architectural investments – investments we are making for future architectures which will pay off beyond the following year.
6. Customer Experience Investments – investments to fix negative customer usability.
7. NPI (New Product Introduction) - cross functional projects which are on the NPI roadmap.
8. Industry Disruptive Events (what are our best ideas to disrupt the industry?)
9. Areas we should be worried about but aren’t today (where should the dog be barking?)
10. Where applicable, investments relating to utilization, efficiency, latency, availability, Move-to-AWS, and information security.

• Appendices:

1. FAQ
2. Operating Expenses and Headcount
3. Summary of Incremental Headcount Request
4. Key Accomplishments and Misses
5. Metrics and Goals for Next Year

Guidance:

• Don't wait till OP1 to introduce a big investment area; review a PRFAQ several months in advance
• Get guidance or test investment appetite before baking plan
• Be vocally self-critical - it builds trust
• Make every word count
• Beware the kitchen sink - cover fewer topics better rather than more topics thinly
• Move less important topics to back-up documents
• The primary beneficiary of the doc is you and your team, to clarify your thoughts - make sure you love your plan
• During review - listen deeply to learn, admit mistakes and when you don't know the answer

OP1 Writing tips

The below is guidance only. You should ask your organization for expectations of your OP1 documents.

1.Goal Writing

Providing Rate Improvements. The structure should cover:

From A (bps|percent) of (Units|GMS|orders) (WW|Country-list|Region) in (Date-Range-Relevant) to B (bps|percent) of (Units|GMS|orders) (WW|Country-list|Region) in (Date-Range-Relevant). An improvement of XX% (Range-Compare-Relevant) against a goal of C (bps|percent) of (Units|GMS|orders) WW in (Date-Range-Relevant).

Example:
We went from producing 200 units of donkey-feed WW in FY2012 to 300 units in FY2013, an improvement of 50% YoY against a FY2013 WW goal of 250 units (a 25% improvement). We exceeded the goal by speeding up the delivery of the donkey-feed-o-matic to March from its original delivery date of June 2013.

Whenever possible, use full year to full year comparisons when possible. If that is not possible, for example it is mid-year and you are writing an OP1, then compare the current year forecast to the prior year actuals:

Example:
In FY2012 we produced 200 units of donkey-feed WW. YTD we have produced 224 units of donkey-feed and are green on our FY2013 WW S-Team goal of 250 units. We project that for FY2013 WW, we will deliver 290 units, exceeding our S-Team goal by 40 units or 16%. For FY2014 WW we will take a more aggressive goal to deliver 350 units, a 20% increase, leveraging the success of the donkey-feed-o-matic.

2.Highlights

Three things are needed in the highlights section:

• What went well?
• How did it exceed some goal? (If there's no goal, this should be BIG.)
• How are we maximizing the good in the future (double-down, more investment, etc)? Some teams put in a section for each highlight: "lessons learned" with the take-aways from the positive surprises.

Example:
Donkey-Feed-O-Matic Launches Ahead of Schedule, Exceeds Production Goals:
In March 2013 we activated the Donkey-Feed-O-Matic in production mode, three months ahead of schedule. The Feed-O-Matic at peak currently produces 14 micro-units per hour, up from the 8 micro-units of the Feed-o-Tron. It is too early to tell the final expected production as the ops team is still ramping, however we are optimistic for a peak of 18 micro-units per hour by November 2013. This was achieved by optimizing the delivery schedule by performing our first Feed-O-Matic Kaizen in January 2013, two months prior to the installation of the hardware. Donkey ACES team pioneered the concept of pre-Kaizen by role-playing the installation without the actual components; this resulted in critical factory layout changes prior to installation. We have written training materials and demonstrated the pre-Kaizen concept with the Amazon ACES program leaders, the first pre-Kaizen with the Donkey-bedding ACES team begins in September 2013.

3.Lowlights

Three things are needed in the Lowlights section:

• What went badly / what goal is impacted? (dates, numbers, give details)
• What impact did it have to our customers / business?
• What are you doing or going to do differently? Make sure the change is at the right scope for the impact. Will there be a re-org? Will extra monitoring be put in place? Expect this to be questioned.

Example:
Manual Donkey Feed Machine Damaged in Repair Attempt:
In March 2012, the Feed-o-Tron, the manual feed manufacturing system was damaged during a critical repair to it's power supply. As a result, 23 units of donkey-feed were destroyed, delaying the delivery of feed to twelve customers. The affected customers have been compensated $XXk for the missed delivery plus penalties. In our COE we determined that the short term fixes and monitoring improvements made on the Feed-o-tron are not sufficient to maintain successful operations. Therefore in 2013, we will replace the Feed-o-Tron with the next generation model, the Donkey-Feed-O-Matic which uses a more secure power supply system and provides better unit production per hour. It is scheduled to go online in June 2013.
Lesson Learned:
While operations personnel had complained about the zapping noises coming from the power supply, their concerns were dismissed as rare anecdotes. We have since instituted a ticketing system for these complaints so we can quantify and track them. Also, requirements for the Feed-O-Matic were modified to include an emergency shut-off button near the power-supply for safety concerns.

4.Themes

Some departments / portions of the company utilize themes to act as the bridge between tenets and initiatives. Themes must set the tone for a collection of initiatives and cover the strategic thinking that informs the choice of work items that are worthy of inclusion for leadership review. Themes often tie into the three-year (3Y) update (some OP1s have a 1/2 page on a recap of the 3Y mission, with Year 1 being set at 365 days in the future of the document's intended review date, Year 2, Year 3, respectively). With a large org, adding a percentage of headcount associated with the themes gives a great macro-perspective of how the leadership for that org thinks about the space and has prioritized. We don't recommend more than 5 themes as it starts to get too hard to track, however as your org size increases, this may increase too.

Themes must accomplish a few goals:

• An explanation of the challenge that cannot be met within the one-year (1Y) time frame. This is usually an aspiration goal reminder.
• Direct statements of the choices we made (don't be vague). You'll write a lot of "We will do X", and "In year one, we will improve recall without impacting precision".
• Serve as the overarching linkage between multiple initiatives. In other words, key initiatives map many: 1 to a theme. "Operations Burden Reduction" / "Platform Investment" / "Developer Facing Features".

5.Key Initiatives

Five things needed in each paragraph / topic:

• What's the problem?
• What metric do you use to measure success (new metric / existing metric)? Describe that you are aware of how customers are going to experience this.
• How are you going to do this?
• Why is this important to the business?
• Goal (SMART goals only).

Example:
Lasers, Pew Pew [2HC][Q3 2014]:
Red Dwarf achieved 99.3% more processing uptime in FY 2018, missing our goal of 99.9%. To meet our 2019 goal, we need to remove 685 hours of downtime per year. The dominant factor of downtime is asteroid strikes monitored by video scans. We have found that asteroid strikes represent 612 hours of downtime per year and may be easily solved. In 2019, we will launch satellite laser cannons into space to vaporize asteroids before they hit. Our lasers are estimated to reduce asteroid hits from an average of 16 hits / month to 3 hits / month, eliminating 42% of the repair costs to Red Dwarf of $100MM each year, this project will save $42MM in FY2019, more than offsetting the one-time cost of $3MM for lasers and the incremental $1MM for yearly laser maintenance.

6.Headcount

Sometimes referred to as "Staffing Elasticity", "Headcount Recommendation", "Funding", "Budget" etc. the headcount discussion (usually towards the end of the doc) makes some sort of recommendation and informs how the program believes its funding is positioned for the next year. The following items are encouraged:

• What is your current headcount and makeup (job families, number of headcount, levels, locations)?
• What happens if funding remains the same? This is the above / below the line bottoms up project name -> headcount cost spreadsheet you'll attach as an appendix.
• What happens if funding drops by 10%? What projects are lowest on the priority list and what you would cut?
• What do you recommend funding be and why?

Example:
Red Dwarf Software Staffing Elasticity:
Currently this program is funded with 15 HC, (2 SDMs, 8 SDEs, 1 TPM, 2 SDET, 2 QAE). If we reduced staffing by 10%, we would change the following commitments in partnership with our customers (team X and team Y). (1) Quarterly painting of Red Dwarf would reduce to yearly re-painting. (2) Development of charging system for the Kryton series robots. (3) Improved filtering algorithms for the selection of iron-rich asteroids vs silicon. We recommend adding 7 HC (5 SDEs, 1 TPM, 1 QAE) as mentioned in the 2019 Operations Plan (see Appendix C for the Resource Allocation table). This increase in funding will result in 500,000 additional tons of iron production by 2019 and improved employee morale and productivity.

7.Critical Evaluation Questions

When reviewers are looking through your OP1, they may have the following questions that you would want to address. Here's some examples:

• Does your mission -> tenets -> strategic themes -> key initiatives -> goals -> headcount-decisions flow together?
• Does your plan reflect the prioritization of your themes and tenets?
• How does this doc compare against last year's OP1 / OP2? (Sometimes OP1 docs have a "this year so far" section with work-in-progress towards last OP1/2's goals, are you changing course?)
• For each of your misses, do I see progress in your key initiatives to address the miss and prevent it from repeating? Do you have a COE link? Put it in a footnote if you do.
• If your team is Tier-1, is there a list of risks to the systems called out with things you're worried about and plans to fix? Some teams have a section on risks to the plan, and risks to the business if the plan doesn't finish.
• For your funding request, are you calling out non-recurring-expenses budget and highlighting the non-headcount related costs? Make a spreadsheet in an appendix.
• For your funding request, do you demonstrate efficiency metrics and gains for OPEX (operating expense) costs? Are you growing linearly with the business on OPEX? Are you scaling?
• For your goal statements, do they align with S-Team objectives?
• For your technical-debt / operational-pain / ticket-reduction challenges, are you properly funding those to improve your efficiency next year? Don't complain; solve.
• How have you made your choice about Headcount recommendations? Why are you requesting funding for an additional XX SDEs and a YY SDETs and not more/less/other-job-families?

Answering as many of these as possible in the document, may reduce the amount of time you spend discussing the DOCUMENT. Instead, the discussion will be what the team might do next year and/or how fast the company is moving in this space. This is the best kind of discussion to have with a large audience of reviewers.