HomeServicesCloud Cost Optimisation

Cloud Cost Optimisation in New Zealand

Cloud bills grow the way subscriptions do: nobody decides to spend more, and nobody is watching the total. A review is one of the few engagements that reliably pays for itself in a quarter.

$12k+

Typical project value

2 min

To post a project

What do you need built?

Start here and add detail as you go.

Free - No obligation - Your details stay private

What people usually mean by cloud cost optimisation

A cloud cost brief is a bill that has been climbing without anyone approving it. The work is measurement, then a prioritised list of changes with a dollar figure and an effort estimate against each. Most of the savings are unglamorous, and most of them are available within a month.

A review

Someone reads your bill and your environment and hands you a ranked list of savings with the effort against each. The usual and best first step.

Review and implement

The same, then the changes are made — resizing, scheduling, cleaning up and committing to discounts where the workload is stable.

Cost governance

Tagging, budgets, alerts and per-team reporting so the bill stays explicable without another review next year.

What it costs

Indicative ranges from projects posted here. Use them to sense-check a quote, not to budget precisely.

Cost review

Assessment of what you run and what it costs, with a prioritised list of savings.

$6k – $15k

Review and remediation

The above, with the changes implemented and verified against the following month's bill.

$15k – $45k

Governance and architecture

Tagging, budgets, alerting, per-team reporting and design changes that reduce cost structurally.

$45k+

Questions worth asking a provider

·What do you expect to save, and how confident are you before you have looked?

·How are you paid — a fixed fee, or a share of the savings?

·Will the list distinguish savings that need no change from those that need engineering?

·Will you verify the savings against next month's bill?

·What will you leave behind so we can see the bill by project or team?

Common questions

How much does a cloud cost review cost in New Zealand?

A review producing a prioritised list of savings is $6,000 to $15,000. Implementing the changes and verifying them against the following month's bill is $15,000 to $45,000. Adding governance — tagging, budgets, alerts and per-team reporting — starts around $45,000. On a bill of a few thousand dollars a month, the review usually pays for itself within a quarter.

What do these reviews typically find?

The same handful of things almost every time: instances and virtual machines sized for a peak that never arrives; test and staging environments running overnight and at weekends; storage, disks and snapshots orphaned when something was deleted; committed-use discounts not taken on workloads unchanged for two years; data transfer that a design change removes; and services someone enabled for a trial and forgot. None of it is clever and together it is frequently a third of the bill.

Should we pay a share of the savings?

Be careful. It aligns the consultant with finding savings, which is good, and it also rewards short-term cuts that create problems later — undersizing production, reducing backup retention, or removing redundancy you were right to have. If you do use that model, agree in advance that changes affecting availability or backups are excluded, and that savings are measured on the actual bill rather than on a projection.

Will this make anything less reliable?

It should not, and this is the question to ask directly. A good review separates savings that carry no risk — deleting orphaned disks, switching off idle test environments — from those that involve a trade, such as reducing redundancy or shortening backup retention. You should be shown that distinction and asked to decide, rather than discovering later that resilience was traded for a lower number.

How do we stop the bill creeping back?

Tagging and budgets. Tags let the bill be read by project or team rather than as one number, which is what turns "cloud costs $8,000 a month" into a decision someone can make. Budget alerts at thresholds you choose, going to a person who will act, catch the next drift in weeks rather than at the annual review. Both are cheap to set up during work that is already happening.

Cloud Cost Optimisation projects open right now

Titles and categories are public. Full briefs, budgets, regions and buyer detail are visible to approved providers only.

Updated daily

No open projects in this category right now. New briefs are posted every week.

Spend an hour on it yourself first

Both AWS and Azure give you the tools free, and an hour with them makes every quote you receive better — and occasionally makes the engagement unnecessary.

Open Cost Explorer or Azure Cost Management and look at:

  • The last twelve months as a graph. When did it start climbing?
  • The top five services by spend. Is anything there a surprise?
  • The split between production and everything else.
  • Anything with a name nobody recognises.

Take those figures into the conversation. A consultant who can see the shape of the problem before starting will give you a far more useful proposal.

What a review always finds

The savings are unglamorous and consistent. In rough order of how often they turn up:

  1. Oversized compute. Machines sized for a launch-day peak that never came, running at eight percent utilisation for two years.
  2. Non-production running all night. Test and staging environments cost the same at 3am on Sunday as at 10am on Tuesday. Scheduling them off is often a quarter of the bill.
  3. Orphaned storage. Disks left behind when a machine was deleted, snapshots from a migration in 2023, backups with no retention policy.
  4. No committed-use discounts. Workloads that have run unchanged for two years still paying on-demand rates, when a one-year commitment would cut them substantially.
  5. Data transfer. Traffic between regions or availability zones that a design change removes.
  6. Forgotten services. Something enabled for a trial, still billing.

None of it requires cleverness. Together it is frequently a third of the bill.

Insist the list distinguishes risk

This is the difference between a good review and a dangerous one.

Free savings — no trade-off at all. Deleting orphaned disks, switching off environments nobody uses, removing a service nothing calls. Do all of these.

Engineering savings — real work, real payback. Resizing after measuring, caching to reduce database load, changing a design that shifts data between regions.

Trades — cheaper and less safe. Reducing redundancy, shortening backup retention, running production without a standby, dropping a support plan.

You should be shown which is which and asked to decide the third category yourself. A report that presents all three as "savings" is one to be careful with.

Be thoughtful about payment on results

Some firms offer to take a share of what they save. It aligns them with finding savings, which is genuinely useful, and it also rewards the third category above.

If you use that model, agree in writing:

  • Changes affecting availability, redundancy or backup retention are excluded.
  • Savings are measured against the actual bill, not a projection.
  • The share is capped, and time-limited.
  • Committed-use discounts count once, not every month for a year.

A fixed fee is simpler and usually cheaper on a bill of a few thousand dollars a month.

Commitments: right size, right length

Savings Plans, Reserved Instances and Azure Reservations cut rates substantially in exchange for committing for a year or three.

They are the largest single saving available, and the trap is committing to capacity you are about to stop needing. Sensible practice:

  • Commit only to the baseline you have run for the last twelve months, not to the peak.
  • Prefer one year over three unless the workload is genuinely settled.
  • Right-size before committing — a commitment to an oversized machine locks in the waste.
  • Know what happens if you need to change.

Tag everything, then set budgets

The reason cloud bills drift is that nobody can read them. "AWS costs us $8,000 a month" is not a fact anyone can act on. "The reporting environment costs $1,900 a month" is a decision.

Ask for tagging as a deliverable — every resource tagged with a project, an environment and an owner — and for budget alerts at thresholds you choose, going to a named person.

That is what stops you buying this engagement again in eighteen months.

Verify against the next bill

Make this part of the agreement. Projected savings and realised savings are different numbers, and only one of them appears on your credit card.

Ask for a short follow-up after the next full billing cycle comparing what was expected with what happened, and explaining any difference. Firms confident in their work will offer it.

Comparing five quotes

Read for:

  • Who asked to see your bill before quoting?
  • Who separated risk-free savings from trades?
  • Who offered to verify against the following month's bill?
  • Who included tagging and budget alerts?
  • Who was clear about how they are paid?

Describe what you run, roughly what you spend, when it started climbing, and whether anything is off limits. Consultants who do this regularly will ask for read access to your billing data before they promise anything.

Related services

Not quite what you are after? These are the briefs most often confused with cloud cost optimisation.

Get up to 5 quotes for your cloud cost optimisation project

You choose who may contact you. Free to use, no obligation.

Start your project