How to Build a 5-Year Plan for Your Company - 5 Lessons from China's Playbook

Lesson 1 — Few and memorable beats complete and unread

China uses 20 headline indicators for the entire country, across five groups: three for the economy, three for innovation, seven for living standards, five for green and low-carbon, and two for security.

A simple test for your own organisation: ask a department head mid-meeting what the company is measuring this year. If they answer within ten seconds, the count is workable. If they need to open a document, there are too many — and indicators nobody can recall do not influence daily decisions.

Lesson 2 — Separate "cannot miss" from "would like"

Of China's 20 indicators, 8 are binding (约束性) and must be delivered, while 12 are indicative (预期性), setting direction without obligation. The split is stated explicitly in the document.

Most organisations do not make this split. The result is that everything becomes indicative by default: when not all targets are met, nobody knows which mattered more, and explaining a miss is equally easy in every case.

Type	Character	Organisational examples
Binding	Cannot be missed, named owner, reviewed monthly	Closing the books on time · defect rate below threshold · zero safety incidents
Indicative	Sets direction, adjustable as conditions change	Market share · revenue share from new products · satisfaction scores

A workable rule: no more than five binding indicators. More than that means the decision about what matters most has not actually been made.

Lesson 3 — Do not set fixed numbers for what you cannot control

The 15th Plan contains no growth percentage. It says "kept within a reasonable range" and sets annual figures as conditions require, anchored to a long-run 2035 goal.

The design logic is clear: when you fix a number that depends on external factors, what you get back is often not better performance but better methods of hitting the number — pulling sales into the quarter end, deferring costs across periods, or trimming quality to cut cost.

Lesson 4 — Be willing to change what you measure

The clearest sign of planning maturity is willingness to abandon an existing indicator. This plan moved from "share of good air days" to measuring PM2.5 concentration directly, and from energy use per unit of GDP to the non-fossil share of energy consumption.

Both changes admit the old measure could be satisfied without the underlying problem improving. The same pattern shows up inside companies — counting quotations issued rather than win rate, or tickets closed rather than issues a customer never had to raise twice.

Lesson 5 — Tie every goal to a budget and a data source

China's plan names verifiable projects: roughly 100 zero-carbon industrial parks, over 1,000 km of zero-carbon transport corridors. Targets can be checked against physical reality rather than reports.

At organisational scale, three questions belong on every indicator before approval: who owns it, what does it cost, and which system will the number come from? The last is routinely skipped, then resurfaces as a problem at reporting time.

A five-year plan structure that actually gets used

Condensed into a document an organisation will genuinely use, the following structure is enough for most mid-sized businesses and should not exceed ten pages.

Section	Content	Length
1. Destination	What the organisation will be in five years, as one picture in five lines or fewer	1 page
2. Indicators	No more than 12, with no more than 5 binding, and a named data source for every one	1 page
3. What you will stop doing	Products, customers or processes to discontinue — the part people avoid writing	1 page
4. Major initiatives	No more than seven, each with an owner, a budget and the month first results are due	3–4 pages
5. Risks and responses	No more than five, each with a measurable early warning signal	1–2 pages
6. Review cycle	Numbers reviewed quarterly, assumptions reviewed annually	Half a page

Section three is what separates a real plan from a wish list. A plan that only adds without subtracting spreads resources thinner every year until nothing is done well.

Start this quarter, not at the start of the year

A common mistake is waiting for the fiscal year to begin, when preparing the data takes far longer than writing the plan itself. This table turns the five lessons into work you can start now.

Lesson	Work you can do this quarter	Rough effort
Few and memorable	Put every indicator currently in use on one page and count them	1 week
Separate binding from indicative	Have leadership pick no more than five that truly cannot be missed, then announce them	One meeting
No fixed numbers for the uncontrollable	Identify market-dependent indicators and convert them to ranges with review triggers	2–3 days
Change what you measure	Find indicators always met while the real problem persists, and change the measure	1 week
Tie to budget and data source	Write the source system next to every indicator; flag the ones you cannot complete	2–3 days

That last task has the highest return per hour spent, because the list of indicators with no identifiable source is precisely the list of work that must happen before planning can be real — and it is usually shorter than expected.

Another borrowable idea is what China's plan calls an applied pilot base: test on real work at small scale before rolling out. In planning terms, run a new indicator set in one department for a quarter and check whether the numbers can actually be produced, before announcing it company-wide. We discussed the same principle for AI deployment in part two.

Which organisations should do this, and which should not yet

Organisation profile	Five-year plan?	Reason
Investments with payback beyond three years — machinery, plant, core systems	Strongly suitable	The decision horizon already exceeds the fiscal year
Agencies reporting against master plan or public sector indicators	Strongly suitable	Must align with national planning cycles regardless
Businesses shifting product or market structure	Suitable	Need a document stating clearly what will be discontinued
Businesses with unstable cash flow	Not yet	Build a 12-month plan that controls cash first, then extend the horizon
Organisations without a reliable central data source	Possible, but unmeasurable	Do the data work alongside the plan, not afterwards

The review cycle — the part people forget to write down

A plan with no scheduled review only gets reviewed once something has already gone wrong, which is the worst moment to decide anything because everyone is under pressure.

The workable pattern separates two levels. Quarterly, review the numbers: are we on track? This is quick, because it only covers the agreed indicators. Annually, review the assumptions: is what we believed when writing this still true — raw material prices, customer behaviour, new entrants?

China applies the same principle nationally: annual GDP figures set as conditions require, with the full assumption set refreshed every five years, and indicators genuinely changed when that review happens. Consistency of cycle matters more than frequency, because everyone knows in advance when the conversation will occur and can prepare data for it.

One final suggestion: write into the plan the conditions that trigger an early review — sales below the stated range for two consecutive quarters, or a key input cost moving beyond a defined threshold. Writing triggers in advance makes a mid-course review a normal part of the rules rather than a signal that something has gone wrong.

Data is the precondition for measuring anything

Follow all five lessons and they converge on one point: the organisation needs real numbers it can produce on demand. In practice, the best plan is meaningless if the review meeting still waits three weeks for departments to compile figures.

An ERP system fills that role directly — the single place where cost, inventory, budget, procurement and disbursement data meet, with every line traceable back to a source document. For public sector agencies reporting against master plan indicators, the visible benefit is shorter reporting cycles and fewer arguments about whether the numbers are right. The underlying integration challenge is covered in how to make every ERP system communicate.

On Saeree ERP's side, we will say plainly that a system does not improve a plan by itself. Deciding what to measure remains an executive job. What the system does is make what you chose to measure genuinely measurable, and auditable line by line.

Conclusion

The five lessons reduce to one sentence: keep indicators few enough to remember, clear enough to know what cannot be missed, flexible where you lack control, changeable when they measure the wrong thing, and tied to a budget and a system that can produce the number.

This series began by unpacking the 15th Five-Year Plan, moved through the AI+ initiative, technology self-reliance, a comparison with Thailand's national strategy, and the price pressure businesses face. The conclusion running through every part is the same: whatever you are responding to, it starts with knowing your own numbers.

Plans do not fail because the goals were wrong. They fail because when the review comes around, nobody knows what the real numbers are.

- Sureeraya Limpaibul · Managing Director, Grand Linux Solution Co., Ltd.

References

• NPC of China — Draft Outline of the 15th Five-Year Plan, summary (6 March 2026)
• Xinhua — Draft outline proposes 20 headline indicators (5 March 2026)
• NDRC — Press briefing explaining the draft outline (9 March 2026)
• National Public Complaints Administration — New indicators in the 15th Five-Year Plan draft (8 March 2026)