· M Lokhandwala How a 45-Person Tech Firm Could Recover $42,000 in Annual Margin Without Downgrading Its Software
A practical keep, cancel, or swap framework for identifying SaaS waste, costing replacement risk, and recovering operating margin.
An illustrative FoundMargin scenario for finance, operations, and technology leaders
A 45-person software and cloud-consulting firm is growing revenue by 18% a year, but its operating margin is moving in the opposite direction.
Nothing looks dramatically wrong. The company has good clients, a busy delivery team, and a familiar software stack. Yet software subscriptions now cost $4,850 per month, $58,200 per year, across 18 tools.
The problem is not 1 reckless purchase. It is the accumulation of individually reasonable decisions:
- An automation platform moved into a higher usage tier as workflows multiplied.
- An internal-app platform began charging for every support and operations user.
- A customer-support system grew more expensive as agents and channels were added.
- Former employees and contractors retained paid accounts after leaving.
Taken together, these costs consume tens of thousands of dollars in annual margin.
This is an illustrative scenario, not a FoundMargin client case study. The figures show how we would examine a software stack, validate the available savings, account for the cost and risk of change, and decide what to keep, cancel, or swap.
The slow accumulation of the SaaS tax
Growing companies rarely design their software stack all at once. They build it under operational pressure.
Sales needs a workflow immediately, so someone creates a Zap. Support needs access to customer records, so the company buys more internal-tool seats. A new service channel is added to the help desk. A contractor leaves, but nobody owns the offboarding checklist.
Each decision solves a real problem. The waste appears later, when pricing tiers, user counts, duplicated functions, and neglected renewals begin to compound.
In this scenario, the annual software spend consists of 4 major areas:
| Cost area | Monthly cost | Annual cost |
|---|---|---|
| Workflow automation | $899 | $10,788 |
| Internal operations applications | $1,500 | $18,000 |
| Customer-support platform | $1,725 | $20,700 |
| Inactive and unnecessary licenses | $726 | $8,712 |
| Total | $4,850 | $58,200 |
Looking only at the $58,200 total would not tell leadership what to do. The useful question is whether each cost belongs in 1 of 3 categories: keep, cancel, or swap.
The keep, cancel, or swap framework
Keep what protects productivity
The objective of a software audit is not to replace as many tools as possible.
Core systems such as email, calendars, company chat, source control, and the primary CRM often carry enormous switching costs. Even when a cheaper alternative exists, migration effort, retraining, disruption, integrations, and employee resistance can overwhelm the license saving.
In this scenario, Google Workspace, Slack, and GitHub remain in place. They are familiar, widely used, and embedded in daily work. Unless usage or contract evidence reveals a specific problem, disturbing them would create more risk than value.
Cancel what has no defensible use
The least risky savings normally come from licenses that should not exist.
Suppose the company finds 12 accounts assigned to former contractors, departed employees, or people who have not used the relevant product in at least 30 days. After checking with the system owners and confirming that the accounts can be removed, those licenses represent $726 per month, $8,712 per year, in hard savings.
This is the strongest kind of finding because it can be supported by invoices, admin-console records, last-login data, employment status, and a confirmed cancellation date. It does not depend on a benchmark or an assumption about future behavior.
Swap only where the full economics work
Some tools are expensive because their pricing scales by user, task, event, or feature tier. Open-source or differently priced alternatives can be attractive in these categories, particularly where the software sits behind the scenes and creates little user-interface disruption.
But a license-price comparison is not enough. A credible decision must include:
- implementation and migration effort
- cloud infrastructure
- monitoring, backups, and upgrades
- security and access controls
- support ownership and incident response
- expected downtime and business impact
- training and process changes
- the cost of reversing the decision if the replacement fails
If the replacement still wins after those costs are included, a swap may be justified. If it does not, the existing SaaS product should stay.
3 possible swap candidates
The following examples use commonly discussed tools to make the economics concrete. They are not universal recommendations, and current pricing and product capabilities must be verified during an actual audit.
1. Workflow automation: Zapier to n8n
Imagine that the company runs 14 automations, including a 6-step lead workflow:
- Receive a web-form submission.
- Enrich the company record.
- Create or update a CRM deal.
- Notify the sales team.
- Record the event in an audit sheet.
- Trigger an internal follow-up.
As task volume grows, the company moves into a plan costing $899 per month. A self-hosted automation platform such as n8n may remove the per-task pricing pressure, but it also introduces infrastructure and support responsibilities.
If the replacement infrastructure costs $20 per month, the simple license-and-hosting difference is:
($899 − $20) × 12 = $10,548 per year.
That is not yet the final saving. Migration effort, testing, maintenance, monitoring, and support must be deducted. The swap should proceed only if the net saving remains material and the company has a credible support owner.
2. Internal applications: Retool to Appsmith
The firm has an internal portal used by 30 support and operations employees. It allows authorized staff to look up customer records, initiate approved refunds, and extend trial periods.
At $50 per user per month, those 30 users cost $1,500 per month, or $18,000 per year.
An open-source internal-app platform such as Appsmith could remove the per-seat license charge. However, the replacement must reproduce the necessary permissions, audit trail, integrations, testing, and support. Access to customer and payment systems makes role-based controls especially important.
The maximum license saving is $18,000 per year. The realised saving is lower after implementation and ongoing operating costs. If the portal is complex, business-critical, or poorly documented, keeping the current platform may be the better decision.
3. Customer support: Zendesk to Chatwoot
15 support representatives handle email, web chat, and messaging inquiries. The current platform costs $1,725 per month, or $20,700 per year.
A platform such as Chatwoot may reduce per-agent fees, but support software is not invisible infrastructure. Agents work in it every day, customers depend on it, and failures directly affect service.
Before recommending a change, the audit would need to verify channel support, reporting, integrations, data migration, retention requirements, mobile access, permissions, uptime expectations, and escalation ownership.
The theoretical license saving is $20,700 per year. The actual saving is whatever remains after the replacement is fully costed. If reliability, missing features, or migration risk erase the benefit, Zendesk stays.
What the financial result could look like
Suppose the gross annual opportunities are:
| Finding | Gross annual opportunity |
|---|---|
| Automation-platform difference | $10,548 |
| Internal-app license difference | $18,000 |
| Support-platform license difference | $20,700 |
| Inactive-account cleanup | $8,712 |
| Gross opportunity | $57,960 |
Now assume the combined annual cost of infrastructure, monitoring, maintenance, and support for the replacement systems is $15,560.
The illustrative net result becomes:
$57,960 − $15,560 = $42,400 in annual margin recovered.
That is the number leadership should evaluate, not the headline claim that open-source licenses cost nothing.
The calculation also needs to distinguish evidence quality:
- Hard savings: confirmed licenses or contracts that can be cancelled.
- Validated savings: costs supported by observed usage, agreed assumptions, and a feasible implementation plan.
- Opportunities: plausible ideas that still require technical, operational, or commercial validation.
Only hard and properly validated findings belong in a savings commitment.
The question every CFO should ask: who fixes it when it breaks?
Open-source software does not eliminate operating responsibility. It relocates it.
For every proposed self-hosted system, leadership should require named answers to the following questions:
- Who monitors availability?
- Who receives an alert when the system fails?
- What is the expected response time?
- How frequently are backups created, and has restoration been tested?
- Who applies security updates?
- How are privileged accounts controlled?
- What happens when an integration or API changes?
- How much internal engineering time is budgeted each year?
- What is the exit plan?
“The community supports it” is not an operating model. The owner may be an internal team, the software vendor, a managed provider, or an implementation partner. What matters is that ownership and cost are explicit before the migration is approved.
How to begin your own software audit
You do not need to begin with a migration. Start by making the current state visible.
1. Build the inventory
Export recurring software charges from accounts payable and company cards for at least the last 90 days. Add contract owners, renewal dates, plan levels, purchased seats, assigned seats, and active users.
2. Validate usage and ownership
Ask the owner of each tool what process it supports, who uses it, which systems depend on it, and what would happen if it disappeared. Admin-console data is better evidence than memory.
3. Find the safest savings first
Prioritize inactive accounts, duplicate products, unused add-ons, incorrect plan levels, and contracts approaching renewal. These usually produce faster and less disruptive savings than platform replacement.
4. Model swaps using total cost of ownership
For any replacement, compare at least 3 years of license fees, infrastructure, implementation, maintenance, support, downtime risk, training, and reversal cost.
5. Rank each finding by evidence, value, effort, and risk
A $3,600 confirmed cancellation that takes an hour may deserve action before a theoretical $20,000 migration that consumes months of engineering time.
The real lesson
The lesson is not that every commercial SaaS tool should be replaced by open source.
It is that software decisions should be revisited once the organization, usage, and pricing have changed. Some systems should be protected. Some licenses should be cancelled immediately. A smaller group may deserve a properly costed replacement.
That is how a company can recover margin without downgrading the way its people work.
All company details, workflows, product comparisons, and financial results in this article are illustrative. Actual recommendations depend on verified contracts, usage, technical requirements, implementation costs, and operational risk.