The real cost of a support SLA for an Indian SMB

Most Indian B2B companies promise a support SLA before anyone has costed it. The promise gets made in a sales call — "four-hour response, guaranteed" — and lands six months later on a founder or ops lead who now needs three people, a shift roster and two subscriptions to keep it. If your buyers are mid-market finance or IT teams who write SLAs into vendor contracts, this is already your problem.

What does a support SLA actually cost per month?

A realistic 12x5, four-hour-response SLA for a business handling 500-700 tickets a month costs roughly ₹2.4-2.8 lakh a month, or ₹30-34 lakh a year, once salaries are loaded with PF, gratuity, payroll overhead and tooling. That assumes Indian metro salaries, not US ones. The money splits into three buckets: people at 75-80%, tooling at 10-15%, and management time nobody budgets for at around 10%. The people bucket is bigger than most founders expect because coverage is a shift problem — you cannot run a 12-hour support window with one person working normal hours, no matter how good the ticketing tool is.

Why does a four-hour SLA cost so much more than next-business-day?

Because response time is a staffing decision, not a software one. A four-hour clock across a 12-hour day means someone is watching the queue at 8pm; next-business-day means the queue can wait until morning. The coverage arithmetic is unforgiving:

• One full-time agent gives about 8 productive hours a day, 5 days a week, minus leave, attrition and training — call it 0.85 FTE of real coverage.
• A 9am-9pm, Monday-to-Friday window needs 2.5-3 FTE per queue.
• Adding Saturday and Sunday coverage pushes the same queue to 4-4.5 FTE.
• A genuine 24x7 commitment needs 5-6 FTE for a single seat, plus a night-shift allowance of 15-25%.

Moving from next-business-day to four hours, 12x5, typically adds ₹90,000-1.1 lakh a month in loaded salary. The helpdesk subscription costs the same either way.

What does it look like as a worked example?

Take a Gurgaon B2B company with 180 active accounts and 600 tickets a month. Its 12x5, four-hour SLA lands at about ₹2.48 lakh a month:

• 3 L1 agents at ₹58,000 loaded cost each = ₹1.74 lakh
• 0.5 FTE of an L2 engineer at ₹1.15 lakh loaded = ₹57,500
• Zoho Desk or Freshdesk at ₹1,200-1,800 per agent = roughly ₹7,000
• WhatsApp Business API, telephony and a status page = roughly ₹9,000

That is about ₹415 per ticket. An outsourced L1 desk quoting ₹350 a ticket looks cheaper until you add 18% GST, TDS under Section 194J, and the L2 escalation you still have to staff internally. In our experience the real saving is 10-20%, not the 40-50% the quote implies.

Should you outsource support or hire for it?

Outsourcing usually wins for L1 volume, night shifts and multilingual queues; in-house wins for anything that needs product judgement. The decision turns on what the quote sheet leaves out, so read it carefully and check whether these are included or billed extra:

• L2 escalation and engineering time
• Weekend and public-holiday coverage
• WhatsApp and voice, not just email tickets
• Ownership of your status page during incidents
• Data handling and retention aligned to the DPDP Act

If your tickets arrive on WhatsApp and email with no shared inbox, the problem is your stack before it is your staffing. That is usually a build problem rather than a hiring one, and fixing it often removes the need for a bigger team at all.

What churn math justifies the upgrade?

An SLA upgrade pays for itself only if it saves more revenue than it costs, and that is a testable claim rather than a judgement call. If moving from next-business-day to four-hour response costs ₹11 lakh a year, it needs to prevent roughly three accounts churning at an annual contract value of ₹4 lakh just to break even. Most B2B teams find that number by segmenting churned accounts on one question: did they raise support tickets in the 90 days before they left? If the answer is no for most of them, your SLA is not the lever — onboarding and product reliability are. Run that cut before you approve the roster, and check the boring stuff first, because an uptime promise is meaningless if your own site is flaky, which a site checker will tell you in two minutes.

Where to start

Start with a 90-day ticket baseline and one segment cut, not a new hire. Pull ticket volume by channel and by account value; you will usually find 10-15% of accounts generating 60% of the volume, and those are the ones an SLA actually protects. Price the roster that covers only them, decide separately whether the long tail can live with next-business-day, and write the exclusions into the contract before sales starts quoting response times again. If you want that modelled against your own numbers first, the team at Maysan Labs typically runs this as a short audit before anyone signs off on headcount.