Accelerate, continue or stop: leading also means deciding what to stop backing

A company doesn't lose focus only by choosing badly. It also loses focus when it avoids choosing at all.

Initiatives that don't fail, but don't demonstrate enough value either. Campaigns that continue simply because they're already under way. Segments that consume energy even though no one can explain their priority. Projects that appear in every meeting without an owner, a date or a closing criterion.

That middle ground is expensive. It eats up calendar time, spreads capacity thin and forces the CEO to keep too many decisions open.

Leading means answering, on an explicit cadence: what should we accelerate, what deserves to continue and what do we need to stop?

Inertia also allocates resources

When a company fails to decide, resources don't stay neutral. They remain tied to whatever started first, has the strongest internal sponsorship, or makes the most noise.

Inertia usually has human causes:

• acknowledging that a hypothesis hasn't been confirmed feels uncomfortable;
• stopping something can be read as failure;
• the sponsor defends the initiative;
• the team fears losing work or influence;
• there always seems to be one more piece of data missing;
• sunk costs weigh more than the future opportunity.

The result is a portfolio with too many priorities. And when everything is a priority, the word no longer gives direction.

What each decision means

Accelerate

There's enough evidence to increase resources, speed or scope. Accelerating doesn't mean enthusiasm; it means the opportunity clears an agreed threshold and the organisation can absorb the investment.

Continue

The hypothesis is still worth pursuing, but you need to learn more. Continuing requires a deadline, the next piece of evidence and limited resources. It's not a way of putting off the decision indefinitely.

Stop

The initiative doesn't justify the capacity it's using right now. It can be closed, archived, or returned to the backlog with explicit conditions for reactivation. Stopping frees up focus and preserves learning.

A fourth category can be useful: reframe. If the pain point is valid but the solution isn't, the company can change the hypothesis without artificially propping up the original project.

Reversible and irreversible decisions

Amazon popularised the distinction between one-way-door and two-way-door decisions. The former are hard to reverse and call for deep analysis. The latter can be undone and should go through a lighter process.

In his 2015 letter to shareholders, Jeff Bezos warned that applying the heavy process for an irreversible decision to every decision slows the organisation down. AWS summarises the model and notes that Amazon encourages acting on reversible decisions with around 70% of the desired information, because waiting for 90% can be too slow. See Elements of Amazon's Day 1 Culture.

This isn't a licence to improvise. It's an invitation to match rigour to the cost of being wrong and to get better at correcting course.

Quality and speed have to go together

McKinsey studied decision-making across organisations and found a strong link between decision quality, speed and business performance. Only 41% of respondents said their organisation's decisions were aligned with strategy and allocated people and financial resources to high-value projects. The full study is available in Decision making in the age of urgency.

Speed isn't achieved by removing analysis. It's achieved by avoiding repetition, clarifying decision rights and escalating only what genuinely needs leadership-level attention.

Organisations that actively reallocate resources tend to be both rigorous and flexible: they keep project and decision cycles in sync and create incentives to move capacity to where it adds most value, according to Admit it, your investments are stuck in neutral.

McKinsey also documents the anonymised case of a retailer that set up a fast decision process to reallocate marketing investment across channels, following real-time customer and performance metrics. By testing and adjusting content in two-week cycles, the company increased its speed to market by 50%. The case appears in How consumer companies are adopting agility. The result doesn't prove that any given cadence always works; it does show what happens when evidence, decision rights and reallocation work to the same rhythm.

Example: fewer, better-resourced projects

Robert Cooper describes the case of a baby-products manufacturer with 16 active projects and resources spread too thin. The leadership team ranked initiatives using attractiveness and productivity criteria, calculated the resources required, and drew a line: projects below it were stopped or paused to concentrate capacity on the best-positioned ones. The case appears in his article Accelerating innovation: lessons from the pandemic.

The article doesn't give a final financial result that would allow the impact to be quantified. Its value lies in showing the mechanism: ranking projects isn't enough; the ranking must be linked to a real resourcing decision.

Stage-Gate also notes that strong reluctance to close projects leads to delays, poor execution and a shortage of resources for the initiatives that matter most. Although it comes from the creator of a commercial framework, the description matches a problem you can see in companies of any size: the portfolio grows faster than capacity.

Criteria before the decision, not after

Without criteria set in advance, every review becomes a political negotiation. A practical system defines:

• the hypothesis the initiative must validate;
• the outcome or signal that would justify accelerating;
• the minimum evidence needed to continue;
• the stopping condition;
• the maximum investment before review;
• the learning timeframe;
• who recommends and who decides;
• where freed-up resources will go.

Criteria can combine customer impact, revenue, margin, learning, strategic alignment, risk, reversibility and capacity required.

Not all of them should carry equal weight. It depends on the type of decision. Consistency means making explicit what matters within each category.

A cadence that turns information into action

Weekly: exceptions and blockers

Reversible operational decisions, obstacles and deviations. These should be resolved close to the people doing the work.

Monthly: performance and learning

Review hypotheses, evidence, resource consumption and the next milestone. Initiatives in "continue" can't stay open without a date.

Quarterly: portfolio and reallocation

Compare initiatives against each other, not only against their original plan. Choose where to increase, maintain or withdraw capacity.

When a critical condition changes

Don't wait for the calendar if regulation, funding, the market, a customer or risk changes. The cadence provides order, not rigidity.

How we help turn the decision into a working system

The sales diagnosis is valuable in its own right: we sort the open sales initiatives, make the cost of sustaining them visible and propose 3 objectives and 3 ninety-day action plans with clear criteria to accelerate, continue or stop. Your team can apply them on its own.

If you decide to continue, depending on the level of support you choose: with oversight we support and review execution and the first real decisions, including the uncomfortable ones; with external sales leadership we also lead the sales effort with objectives, resources and meetings where decisions actually get made. Connecting strategy, finance and operational capacity beyond sales is only part of the work if it's explicitly agreed in the mandate.

The goal isn't to decide for the company: your leadership makes the big decisions. It's for the company to learn to lead itself with a shared view, independent judgement and autonomy.

Questions for the next leadership meeting

• Which initiatives are still active without a testable hypothesis?
• Which project receives resources because of its history rather than its potential?
• Which reversible decision are we treating as irreversible?
• What do we need to observe in order to accelerate?
• What date and evidence put limits on the "continue" category?
• What would we free up if we stopped the lowest-priority 20%?
• Where would we reallocate those resources?

Stopping without reallocating can look like a saving. Stopping and concentrating is leadership.

Conclusion

A company that decides well doesn't always get it right. It spots problems earlier, corrects sooner and stops inertia from becoming the strategy.

Accelerating, continuing or stopping turns a collection of initiatives into a managed portfolio. It protects capacity, reduces noise and gives the team a clear understanding of what matters now.

The question isn't only what more the company can do. It's what decision it's avoiding and what resources remain trapped by that delay.

Clarity to lead. A system to grow.

Frequently asked questions

How should you decide whether to accelerate a project?

There needs to be agreed evidence of value, capacity to absorb resources, and a favourable relationship between impact, risk and opportunity cost.

Does stopping a project mean it has failed?

Not necessarily. It can mean the hypothesis wasn't confirmed, that another opportunity is better, or that the timing isn't right. The learning should be preserved.

What's the difference between continuing and postponing?

Continuing implies a next milestone, resources and a review date. Postponing removes the initiative from execution until an explicit condition is met.

Who should make the decision?

It depends on its reversibility, impact and scope. Operational decisions should stay close to the team; irreversible decisions or significant reallocations require leadership.

References and sources

• 2015 Letter to Shareholders · Amazon
• Elements of Amazon's Day 1 Culture · AWS Executive Insights
• Decision making in the age of urgency · McKinsey & Company
• Admit it, your investments are stuck in neutral · McKinsey & Company
• How consumer companies are adopting agility · McKinsey & Company
• Accelerating innovation: lessons from the pandemic · Cooper, Robert G