Sequencing growth bets across three horizons of the business
Give the core, the adjacent and the new-growth bets separate budgets, scorecards and stop rules, so that the largest business line cannot quietly absorb the rest.

Most growth plans fail on sequence, not ambition. A company funds three kinds of bet from one budget, reviews them in one meeting and judges them by one revenue line. The core business, which pays for everything, takes most of the attention. The new bets get what is left over. We call the method below the three-horizon sequence. The name is ours. It is editorial judgement, not a finding from a published study. We set out to ground it in horizon-planning research and management-consulting studies, and none of the sources we checked address either topic. Treat the method as something to test against your own numbers, not as settled evidence.
Chapter one: Name each horizon by what it is for
The core is the business that pays today's bills. Its job is margin, cash conversion and keeping existing customers. The adjacent horizon holds nearby moves: an existing product sold to new customers, or a new product sold to existing ones. Its job is to produce early evidence and repeat revenue. The new-growth horizon holds the bets that would change the company if they worked. Its job is to produce evidence cheaply, not to produce profit yet. The three jobs call for three different kinds of judgement, which is why one scorecard fails all of them.
The test is simple. Ask the sponsor of each initiative to name its horizon in one sentence. If the sentence needs the words 'and also', the initiative has two jobs and should be split. A typical three-plant manufacturer with one finance lead will find that half its projects are labelled core because they use core equipment, when their real purpose is to win customers in a new market. Relabel them. The label decides which budget, clock and scorecard applies, so it must be right before the money moves.
Chapter two: Give each horizon its own money and its own clock
Set a separate budget line for each horizon and approve the three lines separately. A sensible starting test is to write the intended split down before the year begins, then compare actual spending against it at each quarterly review. If the gap widens without a decision behind it, the split is not being protected. Our judgement is that a fixed percentage works better than a negotiated sum, because a fixed share cannot be argued down in the same meeting where the core asks for one more hire. The percentage itself depends on the company, and we do not offer one.
Each horizon also needs its own clock. The core runs on the annual cycle that the board already knows. Adjacent bets should show first evidence within a period the board can see, and it is reasonable to ask for a read within a year. In our experience, new-growth bets need interim checkpoints that are shorter than their full timeline, so that a failing bet is seen early. These periods are judgement, not benchmarks, and each company should set its own. What matters is that the clock is written down and the sponsor agrees to it.
Funding terms change the economics of a bet, and they change differently by horizon. The World Bank's International Debt Report 2026 says private creditors have returned and bond markets have reopened, while borrowing costs remain at decades-high levels. That report concerns developing-country debt, so it is not a guide to your own rates. Use it as a prompt. Ask finance whether any new-growth bet depends on borrowed money, and what it costs if the rate on that money stays high for the life of the bet. Questions on debt terms, tax treatment of research spending and the legal structure of a venture belong with your accountant or lawyer, not with the strategy paper.
Chapter three: Set the stop rule before the money goes in
Most new-growth bets end in a slow fade rather than a decision. Nobody says stop, so the budget keeps running. The best evidence on this comes from a field experiment with entrepreneurs. NBER working paper 34755, 'Learning to Quit?', by Bailey, Fehder, Floyd, Hochberg and Lee, describes a randomised, multi-year, multi-site study with innovation-driven entrepreneurs. According to NBER, entrepreneurship training aims to raise founders' odds of success. It may also help them recognise when a venture no longer merits further time and resources. The same summary says effective training could shut some startups down sooner while improving outcomes for others.
The lesson for a company is that a well-run stop is a sign that the process is working. The study concerns founders, not divisions inside larger firms, so applying it to internal bets is our judgement. For each new-growth bet, write down two things before the first pound is spent: the evidence that would justify more money, and the evidence that would close the bet. Then set the date on which you will look. A bet that cannot state its closing condition is not yet a bet. It is a hope with a budget line attached.
Chapter four: Measure each horizon separately, including the zeros
Totals hide the horizon that is failing. The U.S. Census Bureau's Annual Business Survey, a joint project of the National Center for Science and Engineering Statistics within the National Science Foundation and the Census Bureau, measures owner demographics, R&D and innovation among US businesses. Its stated aims are better data quality and less respondent burden. Its design records a firm with no R&D as $0 and then skips the R&D follow-up questions. A zero of that kind says nothing about whether the firm tried something and stopped, or never tried at all.
Apply the same discipline inside your company. Record 'no spend' differently from 'spent and stopped', and keep the learning from every stopped bet in a place the next sponsor can find. Then report each horizon to the board on its own page: spend against plan, outcomes against the closing and continuing conditions, and every stop since the last meeting. If one page is empty, that is information. It may mean the horizon is not being fed, which is a different problem from one that is being fed and failing.
Is the sequence working?
Answer each question for the latest budget round. Any 'no' is a decision to take before the next round.
- Does every initiative carry one stated horizon?
- Does each horizon have its own budget line, approved separately?
- Is the core budget capped, so it cannot absorb new-growth money by default?
- Does each adjacent and new-growth bet have a written closing condition and a review date?
- Are stopped bets recorded, with what the company learned from each?
- Does the board receive spend and outcomes for each horizon on separate pages?
- Has finance checked whether the funding terms of any bet change its economics?
Outgrace checklist
Where it breaks
The method breaks in predictable places. The first is pressure on the core. When margins tighten, the core asks for the new-growth budget, and it usually gets it, because the core's case is easier to read. The cap has to be enforced by someone who does not own any of the three horizons, typically the chair or a board committee. The second is the adjacent horizon, which is often treated either as a small core project with no patience or as a new-growth bet with too much. Set its clock explicitly and check it each quarter.
The third failure is the headline number. NBER working paper 35551, 'Chinese Sputnik Moments?', reports the headline that China's patent filings are about three times US filings. Critics, cited in the same NBER material, argue that the patent count overstates real innovation. They say incentive schemes produce many incremental filings, and that Chinese progress relies partly on foreign-trained talent and appropriated intellectual property. The lesson for a scorecard is that a count is not an outcome. Pair every activity measure with a measure of what it changed. The adjacent horizon also borrows heavily. NBER's paper 'Foreign Ideas and US Innovation' says 54% of US technologies build on at least one foreign technology. Record what each adjacent bet draws on, and where it came from, as part of its scorecard.
Start on Monday with the list, not the budget. Ask every sponsor to name a horizon in one sentence, and move any initiative with two answers into two rows. Next, write the split you intend for the year and send it to the board before the budget meeting, so the board sees the intention first. For each new-growth bet, write the closing condition and the review date, and file them where the chair can see them. Last, ask finance to produce the first three-page board view: one page per horizon, plus a list of stops since the last meeting. The first version will be rough. The point is that the gaps become visible, and that the company can see where its effort and capital actually went.

