"Enter a new market in twelve months." How do you turn that goal into engineering bets with a sequence and kill criteria?
Define what entering means, then split the goal into a few engineering bets, each with a hypothesis, a leading indicator and a cost. Sequence the cheapest learning first, release capacity through gates, and agree kill criteria in advance that fire on early signals, not month-twelve revenue.
5 min readUpdated 2026-10-10
0
Answer
Short answer
First define what "enter" means: a first paying customer, a number of live customers or a revenue run rate. Then break the goal into three to five engineering bets, each with a hypothesis, a leading indicator and a cost in engineer-months. Sequence them so the cheapest evidence arrives first. Release capacity in stages, through gates, and agree kill criteria with the CEO before work starts. Kill criteria must trigger on early signals such as pilot sign-ups, because month-twelve revenue arrives too late to stop anything.
Was this answer helpful?
Likely follow-ups
→Sales wants in-region hosting in stage one because one large prospect asked for it. How do you decide?
→A kill criterion fires at month four but the CEO wants to continue. What do you do?
→How do you staff the first stage without stripping the team that runs the core product?
→What would make you recommend not entering the market at all?
"Enter a new market" can mean a press release, one reference customer or a meaningful share of next year's revenue. Each implies a different investment. Ask the CEO which one, and write it down as an outcome: for example, a stated number of paying customers live in the new market within twelve months, with a target sales cycle length.
Also ask what the company already believes about the market and why. If the evidence is one enthusiastic distributor, the first bet is learning, not building. Product leadership usually owns the market case. The questions behind it, such as why growth has stalled and which assets transfer, are the product side of this problem. Your job is to turn the case into engineering bets the company can fund in stages and stop cleanly.
A worked set of engineering bets
Assume a business-to-business invoicing product with about 100 engineers, selling in one country and entering a second. Assume early discovery suggests that customers there expect the local language, local invoice and tax formats, local payment methods and integration with locally popular accounting software. Some prospects also say they need their data hosted in-region. These are case premises, not facts about any real country. The specific invoicing and tax rules come from local advisers and legal review, not from engineering.
Bet
Hypothesis
Leading indicator
Rough cost
Reversibility
1. Localised invoicing
Prospects cannot buy without local formats and language
Pilot customers sending real invoices
10 engineer-months
High
2. Local payment methods
Payment friction is a reason deals stall
Share of pilot invoices paid through the new methods
6 engineer-months
High, if built behind a payment-provider interface
3. Accounting integrations
Integration decides most competitive deals
Deals won or lost with integration cited
12 engineer-months
Medium
4. In-region hosting
A meaningful share of deals require it
Qualified pipeline value that names residency as a blocker
30+ engineer-months plus run cost
Low
The costs are illustrative estimates for this case. The reversibility column drives the sequence. Bet 4 is the most expensive and the hardest to unwind, so it waits until the pipeline shows it is needed, not merely requested.
Sequence and staged investment
Each stage buys the evidence that justifies the next one. Capacity is released at a gate, not committed up front.
Stage 0, weeks 1-8. One team of about six engineers. Sign design-partner prospects, ship localised invoicing for them, and handle payments and accounting exports manually or through a partner. The question is whether customers will pay, not whether the platform scales.
Stage 1, months 3-6. Grow to about fourteen engineers. Complete local payments and the most requested integration. Measure pilot activation and sales cycle length.
Stage 2, months 7-12. Grow to about twenty-four engineers if the gate passes. Build further integrations, and in-region hosting only if the pipeline evidence for bet 4 has crossed its threshold.
Loading diagram…
The edge to look at is the one from each gate into the same stop box. It means stopping is a planned outcome at every stage, decided against criteria written in week one, rather than an admission made in month eleven.
Kill criteria, written before anyone is attached
Kill criteria only work if they are agreed while nobody has a stake in the result. Write them with the CEO and the product lead at the start, and name who decides when one fires. Illustrative versions for this case:
Gate 1. If fewer than five of the design-partner prospects have signed a paid pilot by week eight, stop stage 1 and return the team to its previous work.
Gate 2. If fewer than half of pilot customers are sending invoices weekly by month six, or the sales cycle is far longer than at home, rescope before adding engineers.
Bet 4. Build in-region hosting only if qualified pipeline that names residency as a blocker exceeds an agreed share of the total.
The thresholds are placeholders for numbers the commercial team sets. The structure is the point: early, observable signals, tied to a decision someone has already agreed to make.
Capacity trade-offs in numbers
A new market is never free. In this case, assume 30% of the 100 engineers go on run and unplanned work, leaving about 70 for planned work. Stage 2 at twenty-four engineers takes about a third of planned capacity. State what stops to make room: for example, one home-market feature stream pauses for two quarters and a planned internal tooling project moves to next year. Tell those owners at the start, not at stage 2.
Company context changes how hard you push. A venture-funded startup may accept more risk on a single big bet. A listed company or a private-equity-owned business may want smaller stages and firmer gates, because the cost of a public or reported failure is higher.
Planning for launch day instead of for evidence
The weak answer treats the twelve months as a project plan: build localisation, payments, integrations and hosting in parallel, and launch at month twelve. It spends most of the money before learning anything, and its only kill criterion is the launch result, when the money is already gone. The strong answer has a learning sequence, gates with criteria and a named cost to the rest of the company. It also stays open to a result the CEO did not ask for: the evidence says this market is not worth entering yet.
Fund the next piece of evidence rather than the whole launch, and write the conditions for stopping before anyone is attached to continuing.