Updated 1 Sep 2026 • 4 mins read
Shipped: Side-by-Side Period Comparisons in Explorer
Opslyft Explorer now supports smarter period comparisons: compare any period to the one before it, with like-for-like alignment, partial-period and run-rate handling, and a breakdown of exactly what drove the change, so “is our spend up?” finally gets an answer you can trust.
“Is our spend up this month?” sounds like a simple question. It rarely gets a simple answer, because most comparisons are quietly broken. Compare a full month against one that is only two-thirds finished and spend looks down when it is actually up. Compare a 31-day month to a 28-day one and you are off by 10 percent before you start. The number moves, and nobody can say whether the business changed or the calendar did.
So we fixed the comparison. Today we shipped smarter period comparisons in Explorer, Opslyft's cost exploration view. You can now compare any period against the one before it with proper like-for-like alignment, partial periods handled honestly, and a clear breakdown of exactly what drove the change. “Is spend up?” finally has an answer you can trust, and a reason attached.
What's new Explorer now compares periods intelligently. Pick a range, this month, last week, a custom window, and see it against the prior period on a like-for-like basis. Partial periods are handled with run-rate projection so you are never comparing a half-finished month to a whole one, and every comparison breaks the change down by service, team, and model, so you see not just that spend moved but why.
The Problem With Naive Period Comparisons
Cost tools love to show a big percentage next to last period, and that number is often misleading. Three things break it: partial periods (today's month-to-date against last month's full total), mismatched lengths (28 versus 31 days, or four versus five weekends), and no explanation (a number went up, but which team, service, or model moved it?). Acting on a broken comparison is worse than not comparing at all, because it sends you chasing a change that was really just the calendar. Good comparison is the foundation of good cloud cost forecasting, and it starts with an honest baseline.
What Shipped
Smarter period comparisons in Explorer give you:
- Preset and custom periods. Compare this month, last week, this quarter, or any custom range against the period immediately before it, in a click.
- Like-for-like alignment. Comparisons account for period length and shape, so you are measuring a real change in spend, not a difference in the number of days.
- Partial-period handling. When a period is still in progress, Explorer uses run-rate projection so month-to-date is compared fairly, not against a completed month.
- Driver breakdown. Every comparison shows what moved: the services, teams, and models behind the change, so the delta comes with a cause, not just a number.
What You Can Do With It
- Answer “is spend up, and why?” in one view, with the drivers attached.
- Run a trustworthy month-over-month or week-over-week review without exporting to a spreadsheet.
- Separate real cost changes from calendar effects before you raise an alarm.
- Trace a period-over-period jump straight to the team, service, or model responsible.
Because the breakdown ties each change to an owner, it builds on the same cost allocation that maps every dollar, including AI spend, to a team, so a comparison is never just a number without a name.
How to Use It
It takes a few seconds and no setup.
1. Open Explorer. Go to Explorer and select the spend you want to look at, all-up, or filtered to a team, service, or model.
2. Pick a period. Choose a preset like this month or last week, or set a custom range.
3. Turn on comparison. Compare against the prior period; Explorer aligns it like-for-like and projects any partial period on a run-rate basis.
4. Read the drivers. Expand the change to see which services, teams, and models moved it, and act on the real cause.
Why It Matters
A cost comparison is only useful if you can trust it, and until you can, every review starts with an argument about whether the number is even real. Honest, like-for-like period comparison ends that argument. It turns the monthly cost review from a debate about the calendar into a conversation about the business, and it turns a scary percentage into a specific, ownable change you can actually do something about. Small feature, but it is the difference between a chart people distrust and one they act on.
Get Started
Smarter period comparisons are live in Explorer now. If you already use Opslyft, open Explorer and turn on comparison on your next cost review. If you do not, book a demo or explore cost visibility to see your cloud and AI spend, compared honestly and explained, in one place. We will keep shipping the things that make “what changed, and why?” a question you can answer in seconds.
FAQs
A new capability in Opslyft Explorer that compares any spending period against the one before it on a like-for-like basis, handles partial periods with run-rate projection, and breaks down exactly which services, teams, and models drove the change.
Because they often compare mismatched periods: a partial month-to-date against a full month, or a 28-day month against a 31-day one. That makes spend look up or down for reasons that are really just the calendar, not a real change in the business.
When a period is still in progress, Explorer projects it on a run rate rather than comparing an incomplete period against a completed one, so month-to-date spend is measured fairly against the prior period instead of looking artificially low.
Yes. Every comparison breaks the change down by service, team, and model, so you see not just that spend moved but which parts of your estate moved it, and who owns them.
Open Explorer, select the spend you want, pick a preset or custom period, and turn on comparison against the prior period. Explorer aligns the periods like-for-like, projects any partial period, and shows the drivers behind the change.
Yes. Explorer compares cloud and AI spend together, so you can run like-for-like period comparisons across AWS, Azure, and GCP alongside token, GPU, and model costs, with the drivers attributed to the responsible team.