The Stack Grew While You Were Not Looking
I counted the SaaS invoice line items for the team I manage every January. Last year I got to forty-seven. The year before that, thirty-eight. Two years ago it was twenty-two. Nobody on the team signed up for forty-seven subscriptions on purpose. Each hire brought a favorite project tracker. Each client added a video tool of their own. Each AI pilot spawned a new vendor line and a new admin panel, and the invoices kept stacking up while the team got less coherent about where the real answer lived.
McKinsey State of AI 2026, built on interviews with 1,700+ executives across 97 markets, put a number on the pain. Forty percent of IT executives said AI tool proliferation was already harming collaboration and efficiency. Another 32% said they decided not to buy software because they could build it in house with agentic coding tools. That second number is the interesting one. Half of the executive cohort is already treating software procurement as a solve-for-later decision, and the other half is feeling the cost of the tools they still bought.
I have been building and running remote stacks for eight years. Sprawl is not a buying problem, it is a merge problem. Buying another tool will never fix it, because the fix is the opposite of buying: you pick three things to keep, you write a migration plan for the rest, and you give each one a date.
The Three Axes of Consolidation
Consolidation only works if you have an axis. Without an axis, every cleanup is just a reshuffle. I use three.
Axis 1: Single Identity Provider
Every SaaS logon goes through one identity layer. One SSO. One SCIM for provisioning. One directory. When a person leaves, one deactivation clears every system. When a new hire starts, one workflow opens every account. This alone removes the biggest security surface a distributed team carries, because stale accounts are the single most common breach vector I have seen in a post-incident review.
Anything that requires a separate password, a separate group, or a separate admin flow is a candidate for the retirement queue. It may end up kept, but it needs a very good reason.
Axis 2: Single Source of Truth per Domain
Every domain has exactly one canonical place. Docs, tasks, data, decisions, and messages. If two tools write to the same domain, one of them is a duplicate wearing a license cost. The one you keep is usually the tool the team actually searches when they need a real answer, not the one that sounds better in a demo.
Axis 3: Single Admin Panel
One place to see renewals, spend, access, and deactivations. If your admin work happens in twelve dashboards across four credit cards, you do not have an admin panel, you have twelve surprise renewals waiting to happen. Consolidation along this axis is what makes the renewal leverage of the next section possible.
The Keep vs Kill Score
Score every vendor against four questions.
Does the tool own a source of truth for a domain the team cares about? Did the tool pass your security review? Does the tool integrate cleanly with the other two axes? Is the team still using it daily?
A clean yes on any single question is a keep. A clean no on all four is a kill. Anything in between is a keep for now, but write down the reason and revisit it at the next review.
The table below compresses the six decision areas into a single keep-kill reference you can actually use at the review meeting:
| Decision area | Keep the tool | Kill the tool |
|---|---|---|
| Data ownership | Owns the canonical source of truth for its domain | Stores copies of data that live elsewhere |
| Security posture | Passed your security review, supports SSO and SCIM | No SSO, weak review, stale security posture |
| Integrations | Flows into the identity provider and the admin panel | Siloed, requires manual re-entry, blocked by the other two axes |
| Actual usage | Active daily usage across the team | Under 20% active seats, or no login in 60 days |
| Vendor health | Parent investing, product shipping, roadmap clear | Layoffs, segment spin-off, roadmap going quiet |
| Renewal leverage | Price visible in a single admin panel | Renewal spread across multiple dashboards |
Two of my recent audits are useful illustrations. I kept a niche docs wiki that fewer than twenty people used, because it owned the source of truth and passed security. I killed a general-purpose CRM that two hundred people used, because it did not pass the security review and did not integrate with the identity provider. The keep was the right call even though it felt counterintuitive, because the CRM was one away from losing its last remaining anchor.
The 90-Day Migration Playbook
Every killed tool gets a written migration plan and a hard sunset date. Tools without a date never leave.
Days 1 to 30: Prepare
Write the migration plan. Map which workflows move where. Export the data you need in JSON, CSV, or an API dump. Confirm the destination tool can receive it. Freeze new seats on the dying tool immediately, because new seats make the migration window longer, not shorter.
Days 31 to 60: Dual Run
Run both tools in parallel. Migrate live workflows in chunks. Train the team on the destination tool during this window, not before, because training before migration creates two habits instead of one. Watch for the first real failure, then fix it while you still have a fallback.
Days 61 to 90: Sunset
Cancel the card. Delete the data after confirming the export is complete. Remove the account from the identity provider. Close the ticket. Delete the folder in your drive. Every step after this line is cleanup, and it should feel easy, because you are only removing things.
Supplier-Risk Signals
Consolidation is also a supplier-risk decision. The 2026 SaaS layoff wave is the clearest evidence yet that paying for a product does not mean the parent company is investing in it.
Intuit announced on 2026-05-20 cuts of about 3,000 jobs, roughly 17% of its workforce, and in its FY2026 results announced 2026-08-25 said it will report Mailchimp as a separate segment from fiscal 2027 after scaling back investment in it. Cloudflare cut 1,100 jobs, about 20% of its workforce. Monday.com cut about 630 jobs, roughly 20%, on 2026-09-17. Atlassian cut about 1,600 jobs, around 10%. Wix cut about 1,000 jobs, roughly 20%. ClickUp cut about 22% of staff. All cited AI as the reason.
When an entire cohort of vendors reports the same cut percentage the same month for the same reason, that is not coincidence. That is the industry repricing its own margin, and the repricing is showing up in your renewal notices.
Track four signals per vendor at every renewal:
- Parent strategy shift. Is the parent company explicitly prioritizing another product line? Are the roadmaps for the two diverging?
- Segment spin-off. If the vendor is being split into a separate segment or an independent unit, expect feature freezes and slower support.
- Layoff percentage. Anything above 10% in a single quarter is a leading indicator, because the product roadmap usually tightens the following two quarters.
- Roadmap quietness. Changelogs that go silent for three months are not a bug, they are a signal.
Add these to the vendor scorecard. A tool that scores poorly on supplier risk is a keep only if it also scores strongly on source of truth and integrations. A tool that scores poorly on all three is a kill with priority.
The Renewal Leverage
The reason to run consolidation before the renewal window is not just to save on the tools you kill. It is to see the pricing pressure on the ones you keep, all at once.
After the merge, you have one admin panel, one invoice cycle, and one procurement point. You can negotiate a bundled renewal instead of twelve fragmented ones. You can push back on the per-seat price increase you were silently accepting last year. You can threaten a real alternative, because you now know what would break and what would not.
You also get the honest answer to the question every team asks: how much do we actually spend on software? It is usually less than you thought and less than you want to believe, but it is the only number that matters when a vendor calls with a 22% renewal hike.
The Merge Is the Whole Point
The temptation is to think consolidation means fewer features. It does not. The point is to stop paying twice for the same capability and stop running two tools that fight over the same source of truth.
Pick three things to keep, score the rest against four criteria, write a ninety-day migration plan for the ones you kill, and read supplier-risk signals before every renewal. That is the playbook. It is not glamorous, it does not require a new tool, and it does not need a committee. It needs one person who owns it and a calendar that enforces the sunset date.