Most bad offers in Indian tech are not too low. They are badly built. The number looks fine on a slide and falls apart the moment the candidate opens their calculator, compares the in-hand, weighs the ESOP, and calls their current manager. A fair offer is one the candidate can say yes to without doing mental gymnastics, and one you can defend to finance without overpaying. This is how to get there.
I have made offers that got accepted on the spot and offers that died over a two-lakh gap that a smarter structure would have closed for free. The difference is almost never the headline CTC. It is the shape of the offer and how well you read where the candidate sits against the market.
Start with real bands, not last year's memory
Indian tech comp moves fast, and the bands people quote from memory are usually 18 months stale. Here is a working picture for 2026 across product companies and strong GCCs in Tier 1 cities. Services firms pay lower; top-tier product firms and specialised skills pay higher.
Software engineer fixed-pay bands, Tier 1 (2026)
- 0 to 2 years: 8 to 18 LPA
- 2 to 4 years: 16 to 30 LPA
- 4 to 7 years (senior): 28 to 50 LPA
- 7 to 10 years (lead / staff): 45 to 80 LPA
- Principal / architect: 80 LPA to 1.5 crore total
Treat these as anchors, not truth. A machine learning engineer with real production experience sits at the top of every band. A generalist at a services firm sits at the bottom. The band tells you the room; the candidate's actual profile tells you where in the room to stand.
Fixed versus variable: the in-hand test
Candidates in India increasingly discount variable pay, and they are right to. A 40 LPA offer that is 30 fixed and 10 variable is not a 40 LPA offer to the person receiving it. They have watched variable pay get cut in a bad quarter. So the first rule of a fair offer: keep variable modest for engineering roles, usually 10 to 15 percent of total, and be honest about the payout history.
Joining bonuses are a cleaner lever than inflated variable. If a candidate is leaving unvested stock or a retention bonus on the table, a one-time joining amount that covers the loss closes the gap without permanently raising your band. It is a targeted fix for a specific problem.
ESOPs: the number nobody agrees on
ESOPs are where fair offers go to die because both sides value them differently. The company quotes a notional value at the last round's valuation. The candidate, if they are experienced, values them near zero until there is a liquidity event. Both are being rational.
The fair move is transparency. Show the strike price, the vesting schedule, the last valuation, and whether there is any buyback history. A candidate who has been through a dead-ESOP experience will trust a founder who says "here is the math, here is the risk" far more than one who waves a big notional number. If your cash is tight and your equity is your real pull, sell the equity honestly instead of pretending the cash is competitive. Founders hiring lean should read our take on startup hiring on a budget.
City cost differences are real, so use them fairly
A 30 LPA package in Bengaluru and the same 30 LPA in Coimbatore are not the same life. Rent in central Bengaluru or Gurugram can eat a third of take-home. In Pune, Chennai, Ahmedabad, or Coimbatore the same money goes much further. For remote or Tier 2 roles you can pay 15 to 25 percent below Tier 1 fixed and still make a stronger real offer.
Be careful, though. Candidates know their worth is national now, not local. If you cut too hard on a remote engineer purely on geography, a Bengaluru product firm will simply hire them at full rate. Use city cost to justify a fair number, not to lowball.
Notice periods and counter-offers
The 60 to 90 day notice period is the silent killer of Indian tech hiring. A candidate who accepts today may not join for three months, and in those three months their current employer will counter. Plan for it. Build the relationship through the notice period, keep the hiring manager in contact, and set a joining bonus that vests on day one so leaving is costly to reverse.
Counter-offers usually fail the candidate within a year, but they succeed at the moment of resignation because they are emotional, not rational. The best defense is to have read the candidate correctly before you ever made the offer. If someone is a genuine mover, they will resist the counter. If they were only testing the market, no offer of yours will hold. Reading that difference is exactly what the signals in our piece on signals that predict a yes are built to catch.
Know who is above and below market before you reach out
Here is the mistake that wastes the most recruiter time: you find a great profile, spend three emails building interest, get to the number, and discover the candidate is already paid 20 percent above your band. Dead on arrival. You could have known that up front.
This is where TalentGPT's pay-gap, or comp-fit, signal earns its keep. Before you reach out, it estimates where a candidate likely sits relative to market for their role, level, and location, and it shows the reasoning rather than just a flag. That tells you two useful things. Who is underpaid and therefore movable on money, and who is already above your band so you either stretch or skip. You spend your outreach on people you can actually close.
Paired with the retention read, this changes how you build a shortlist. A candidate who is underpaid relative to market and showing early flight-risk signals is your warmest lead, because they have both the reason and the readiness to move. That is a far better use of a recruiter's day than cold-mailing a list sorted by keyword match.
A fair offer checklist
- Benchmark against a current band, not last year's memory
- Keep variable modest and honest; use a joining bonus for gaps
- Show real ESOP math, or sell equity honestly and move on
- Adjust for city cost without lowballing a national candidate
- Plan for the notice period and the counter-offer from day one
Frequently asked questions
How much of a hike should I offer to move a candidate?
In Indian tech, 25 to 40 percent over current fixed is the common range to move a passive candidate, higher for scarce skills. But hike percentage is the wrong anchor. Benchmark against the market band for the role, not against what the candidate happens to earn today.
Should I count ESOPs in the headline CTC?
Present cash and equity separately. Bundling notional ESOP value into a big CTC number reads as a trick to experienced candidates and erodes trust. Show the equity math on its own and let them value it.
How do I compete with counter-offers?
You mostly win before the resignation, not after. Pick candidates who are genuine movers, stay close through the notice period, and make leaving costly to reverse with a day-one joining bonus and a role they are excited about.
How can I see who is above or below market at scale?
TalentGPT's comp-fit signal flags where a candidate likely sits versus market before you reach out, with the reasoning shown. You can try it in the dashboard and sort your next shortlist by who is actually movable on money.