How to Run a SaaS Rationalization Exercise

The renewal notice lands on Friday afternoon: $96,000 for another year, up 14 percent, auto-renewing in twenty-one days. Procurement forwards it to finance. Finance forwards it to IT. IT knows the platform is installed but cannot tell how many people use it. The department that bought it has a new leader.

Now the company has three weeks to make a decision that should have started three months ago.

I have seen this pattern often. Nobody set out to waste money. A useful tool was purchased for a real problem, the problem changed, ownership moved, and the contract kept renewing. SaaS rationalization is how you replace that drift with a decision.

Some money costs more than it pays.

Build the list from evidence

No single source is complete. Combine accounts payable and card transactions, procurement records, Microsoft or Google identity applications, OAuth grants, browser or endpoint discovery, expense reports, and interviews.

Finance finds tools identity misses. Identity finds free tools finance never sees. Employees identify the platform purchased through a reseller under a different billing name.

Record what matters

For each product, capture business purpose, owner, active users, total annual cost, renewal date, contract term, data stored, integrations, authentication method, administrative owners, backup or export options, and the consequence of losing it.

Naming a department such as marketing does not establish ownership. Name the person accountable for deciding whether the tool remains.

Look for overlap by capability

Products with different names often perform the same work. Group by capability: project management, file sharing, forms, e-signature, automation, analytics, communication, password management, AI assistance, and customer data.

Do not assume two tools are duplicates because their feature lists overlap. One may support a differentiating workflow the other cannot. Ask which processes depend on each and what would have to change.

Separate use from value

Login frequency is evidence, not a verdict. A disaster-recovery platform may be valuable precisely because it is rarely used. A widely used chat tool may create little value if another included platform provides the same capability.

Evaluate business consequence, differentiation, data risk, integration cost, switching effort, and contract timing alongside utilization.

Assign one of five decisions

Keep: the product fits, is owned, and earns its place.

Consolidate: move overlapping work onto an existing platform after validating the gaps.

Renegotiate: the product stays, but licenses, tier, term, or support can change.

Replace: the need is real and the current product is not the right answer.

Retire: the need disappeared, the process changed, or the capability is duplicated.

Every decision needs an owner and date. Otherwise the exercise becomes another inventory.

Protect the exit

Before cancellation, export required data, document retention obligations, preserve records, remove integrations and OAuth grants, update procedures, communicate the change, and assign a short rollback window where practical.

Turning off a subscription without removing access paths leaves invisible risk behind. Moving the data without testing the replacement creates operational risk instead.

Make rationalization continuous

Review new purchases before contract, tier applications by risk, centralize renewal dates, and revisit the portfolio quarterly. The first exercise removes accumulated drift. The operating system prevents it from returning.

The best result is not a dramatic one-time savings slide. It is a company that can explain what software it has, why each product exists, who owns it, what data it holds, and when the decision will be reviewed again.

A 90-person company, by the numbers

Use a representative example. A 90-person company begins with $612,000 in annual SaaS spend across 81 billing records. Once duplicate vendor names and reseller charges are merged, the estate contains 67 products.

The register shows 14 products with no named business owner. Nine renew inside 120 days. Eleven use personal or shared administrative accounts. Six hold customer or employee information without a documented export or retention plan. The company also pays for 127 project-management seats across three platforms even though it has 90 employees.

The obvious temptation is to announce a savings target. Resist it. First determine why the overlap exists.

One project tool supports client delivery and includes a customer portal that contracts require. The second runs software development and connects to the code repository. The third was purchased for a transformation project that ended last year. That third platform can be retired after its records are exported. The first two stay because they perform distinct work.

Next, the team finds 34 premium licenses assigned to employees who use only standard features. Moving 22 people to the appropriate tier saves $13,200 a year. A contract consolidation reduces another $8,400. Retiring the old project platform saves $27,600 after export costs.

The first-year savings total $49,200. The exercise also closes personal admin accounts, removes stale OAuth grants, and gives every remaining product an owner. Those risk changes belong in the result.

Start with a renewal horizon

Build the first working list around time. Products renewing in the next 120 days need decisions first because auto-renewal can remove your use. Flag contracts with notice periods longer than 30 days. A renewal date and a cancellation date are different fields.

Then add annual value and switching effort. A $1,200 product with no sensitive data deserves less meeting time than a $75,000 platform tied to customer delivery. A low-cost tool can still rise in priority when it has broad file access or no accountable owner.

I prefer a visible decision calendar. Each month shows the products entering review, the evidence due, the owner meeting, and the notice deadline. Finance can plan spend. IT can plan access changes. Department leaders have time to test alternatives.

Without that calendar, every renewal becomes an emergency and the incumbent vendor controls the pace.

Interview the work, not the opinion

Asking "Do you use this?" produces unreliable answers. People defend familiar tools and forget occasional ones. Ask an employee to show the last real transaction instead.

Watch where the work begins. Note the record copied from another system, the approval requested in chat, the export repaired in Excel, and the report rebuilt for leadership. Ask what happens when the normal path fails. Ask which customer or regulatory promise depends on the product.

This walk-through changes the conversation. A product with low login volume may support a monthly billing process worth millions. A heavily used product may be a convenient duplicate. Usage is one signal among several.

The interview also reveals training and configuration gaps. A team may pay for an automation platform while using it as a simple form builder. The rational choice could be to downgrade, improve the implementation, or move the form into a product already licensed.

Calculate full change cost

Cancellation savings are easy to overstate. Subtract the cost of export, migration, integration changes, retraining, contract termination, parallel operation, and internal time. Include the risk of losing history or breaking a customer-facing workflow.

Suppose a product costs $36,000 a year and an included platform can replace it. Migration needs 180 internal hours, $8,000 of specialist help, and two months of parallel licensing. At a $60 loaded internal rate, first-year change cost is roughly $24,800. The first-year benefit is modest. The second year looks much better.

That does not make the consolidation wrong. It gives leadership an honest payback period and enough context to sequence the work.

Also calculate the cost of staying. Manual reconciliation, repeated errors, security reviews, weak exports, vendor instability, and poor customer experience may outweigh the subscription price.

Handle data before the cancel button

Every retirement plan needs an answer for records. What must be retained? In which format? Who can read the export? Do attachments come with it? Are audit logs included? How will legal holds or customer requests be handled? Can the company prove deletion after the retention period?

Test the export while the contract is active. Open the files. Restore a sample. Confirm that dates, owners, comments, and attachments remain usable. A ZIP file nobody has examined is only a hope.

Then remove the paths around the product. Revoke single sign-on and OAuth grants. Rotate integration credentials. Disable service accounts. Remove browser extensions where managed. Update bookmarks, procedures, onboarding, offboarding, and vendor records.

Thirty days later, verify that invoices stopped and no user recreated the account with a company card.

The field mistake: letting the vendor define the need

A renewal conversation often starts with a vendor demonstration of new features. That resets the question from "Does this product still fit?" to "Which tier should we buy?" The frame favors the renewal.

Define the business need before the demo. Write the workflows, measures, gaps, data requirements, security requirements, and exit needs. Then let the incumbent and alternatives respond to the same criteria.

I have watched teams become excited about a new AI feature while the core integration remained broken. The feature may be useful later. It should not distract from the reason the product entered review.

Keep the register alive

After the first cleanup, require a short intake for new software: owner, purpose, users, data, identity method, cost, renewal, integration, success measure, and exit path. The review should take hours for a low-risk tool, not weeks. Fast, proportionate governance gives employees a reason to use it.

Connect onboarding and offboarding to the register. Connect accounts payable to the owner list. Review high-risk and high-cost products quarterly. Review the full estate at least annually.

The mature state is pleasantly boring. Renewals arrive with evidence. Owners know what they are deciding. Data can leave. Duplicate tools need a written reason. Finance and IT use the same list.

That is the point of SaaS rationalization: fewer surprises and better choices, with savings as one result rather than the only headline.

A decision worksheet for each product

Use one short page per material product. Start with the business capability and owner. Add active users, annual cost, contract deadline, data class, integrations, identity method, support burden, and required records. Then write the alternatives and the recommendation.

Include an "evidence still missing" box. The honest answer may be that usage data is unavailable or the contract copy cannot be found. Missing evidence is work, not a reason to guess.

For a keep decision, state the next review and any improvement needed. For consolidation or replacement, state migration cost, parallel period, export test, rollback, and owner. For retirement, state data retention and access removal.

This page makes the decision reusable during budgeting, audits, security reviews, and leadership changes.

Negotiate from the portfolio, not the panic

Vendors know when notice deadlines are close. Begin the review early enough to test tier changes and alternatives.

Bring actual active-user counts. Separate occasional users from daily users. Ask for price protection, shorter terms, ramp schedules, support commitments, data export language, and a clear renewal notice. Remove shelf licenses before asking for a percentage discount.

A lower unit price on licenses nobody uses is still waste.

Common traps

Do not count free products as zero-risk. They may hold data, connect to identity, and become essential without a contract or support path.

Do not let single sign-on create false confidence. SSO can control entry while broad OAuth permissions or unmanaged sharing remain.

Do not merge products because their marketing pages look similar. Trace the actual workflow and customer obligation.

Do not leave the owner field as "IT." Technology can operate a platform. The business owner decides whether the capability earns its place.

Do not declare savings until invoices stop and replacement cost is known.

The meeting should end with dates

Every reviewed product leaves with a decision owner, evidence task, notice date, and execution date. Products awaiting more information receive a short deadline and an explicit temporary status.

The portfolio view then rolls decisions into cash planning, technology work, security cleanup, and employee communication. That is how rationalization becomes an operating discipline instead of a once-a-year spreadsheet.

Start before the next renewal

Choose the largest product renewing more than 90 days from now. Build its decision page, interview two real users, test the export, and compare the alternatives before the vendor meeting.

That single review will expose where contracts, owners, data, and usage evidence live. Fix the process while the stakes are contained. The second product will move faster, and by the time the large portfolio review begins, the company will already have a method it trusts.

Put the result on the next leadership agenda. A completed decision builds momentum better than a large unfinished inventory.