The default rule for repair-vs-replace in the IT world is the "50% of new" heuristic: if repair costs more than half of a new device, replace. It's a useful starting point that turns out to be wrong for most Indian SMEs, most of the time. The rule assumes a market where used hardware has zero residual value, employee productivity isn't being explicitly costed, and the new-device price is fixed and singular. None of those are true for an Indian office in 2026. Used hardware has real value in the buyback and refurbished market. Productive output of an engineer waiting for a repair is a real cost line, not a vibe. And "a new device" comes in five price tiers, three financing options, and a residual-value trade. This post lays out the framework that actually fits.
Why the '50% of new' rule misleads in India
The rule treats a 4-year-old laptop as if it's worth zero on the day a repair quote comes in. In 2026, a working 4-year-old business laptop in India has a real resale market — directly via Cashify, Quikr, and OLX, indirectly via refurbishers who buy in bulk from corporate disposal programs. A clean 4-year-old Dell Latitude i5 in working condition fetches ₹12,000–₹20,000 depending on configuration. That residual value is part of the math.
The rule also treats the new-device cost as a fixed point. In practice, a buyer in 2026 chooses between (a) a brand-new business laptop at ₹55,000–₹95,000, (b) a refurbished similar-class device at ₹25,000–₹40,000 with a 1-year warranty, or (c) a slightly older model still in retail at a discount. The replace cost is a range, not a single number.
And the rule completely ignores employee productivity loss. The repair quote may be ₹12,000, the device may technically be worth fixing — but if the device is going to break again in six months, and each repair event costs the company three days of partially-blocked work from a ₹15 LPA engineer, the productivity cost across two repair events is ₹35,000–₹40,000 of unproductive time, before parts. A clear replacement decision saves that.
The four inputs to the decision
A defensible repair-vs-replace decision uses four inputs, not one. All in rupees.
| Input | What it captures | How to estimate |
|---|---|---|
| Repair cost (R) | What the repair invoice will run | Vendor quote + parts + GST |
| Residual life (L) | How many useful months remain after repair | Device age, prior issues, employee role |
| Productivity loss (P) | What the downtime around this and likely future repairs costs | Employee's daily cost × expected days lost in remaining life |
| Replacement cost (N) | What replacing this device actually costs net of residual value | New device price − resale value of current device |
The repair-or-replace condition is straightforward once you have these four numbers: repair if R + (P × probability of future issues) < N, replace otherwise. Most Indian SMEs estimate R correctly and ignore the other three.
A worked example
A 4-year-old ₹55,000 laptop in a sales rep's hands. Symptom: charging port failure. Vendor quote: ₹8,500 (motherboard charging assembly replacement, labour, GST). Device is otherwise functional, battery health 78%, SSD healthy. The sales rep travels 2-3 days a week.
The naïve "50% of new" rule says repair (₹8,500 is well under 50% of ₹55,000). Now run the four-input check:
- R = ₹8,500
- L = 12–18 months. The device is 4 years old, has had one prior repair (battery replacement at year 3). Probability of another non-trivial repair within 12 months: roughly 30%, based on this fleet's failure curve.
- P = expected downtime cost. Charging-port repair = ~1 day of partial productivity loss (₹3,500 for this employee). Next likely repair within 12 months = another ~2 days of loss (₹7,000). Travel-heavy role makes loss higher than average. Total expected P over remaining life ≈ ₹6,000.
- N = ₹55,000 (new mid-range business laptop) − ₹15,000 (resale value of current device) = ₹40,000.
Decision math: R + P = ₹8,500 + ₹6,000 = ₹14,500. N = ₹40,000. Repair wins decisively. The "50% of new" rule got to the same answer in this case, but by accident — change the inputs slightly (older device, dead motherboard, higher-cost role) and it diverges.
A second example. Same employee, a year later. Symptom: motherboard failure (probable, vendor quote includes diagnosis). Vendor quote: ₹18,000 with parts. Now:
- R = ₹18,000
- L = 6–12 months at best. Device is now 5 years old, with two prior repairs.
- P = high. Probability of another failure in 6 months is approaching 50%; downtime cost in this role is ~₹3,500/day × expected 4 days = ₹14,000.
- N = ₹55,000 − ₹8,000 (lower residual on this older device) = ₹47,000. Or refurbished route: ₹30,000 − ₹8,000 = ₹22,000.
Decision math: R + P = ₹32,000. Replace cost (refurb route) ≈ ₹22,000. Replace, even with the refurbished route. The naïve rule would have repaired — ₹18,000 is still under 50% of ₹55,000 — and would have been wrong.
Refurbished and residual value in 2026 India
The under-discussed lever in this whole framework is the refurbished tier. India's corporate-refurbished market matured significantly between 2022 and 2026; the OEM trade-in programs (Dell, HP, Lenovo Refurbished, and a handful of authorised resellers) now offer business-class laptops at 50–65% of new pricing with a 1-year warranty. For non-frontline roles, refurbished is often the right replacement choice, which changes the N input dramatically.
A practical rule for fleet replacement decisions:
- Frontline / customer-facing / travel-heavy roles: new device, current generation. Reliability and image matter.
- Internal / desk-heavy roles: refurbished, 1-2 generation old. Cost-effective, fine for the work.
- Special roles (design, dev with heavy compute): spec-driven, often new with specific GPU/RAM requirements.
This three-tier replacement strategy halves the average N across a fleet over time.
When to repair even if the math says replace
The framework is a tool, not a rule. Three situations where you repair anyway:
- The device runs niche software that doesn't transfer easily. An old desktop running a specific instrument's control software, an engineering workstation with a discontinued licence keyed to the machine. Repair until that software finds a successor.
- Bridge device. The replacement is being procured but won't arrive for two weeks. A ₹5,000 repair to keep the device alive for that window is justified.
- Specific employee relationship. A senior employee's familiar device. Sometimes the right answer is the comfortable one.
Note what's not on the list: "we've already spent on it". Sunk cost is not a reason to repair.
How a marketplace's diagnosis-first model helps the decision
The biggest input you don't have when making this decision is an accurate R. A vendor estimate based on a phone description is a guess. A pre-repair diagnosis quote — vendor onsite, diagnoses the actual fault, gives a real number for parts and labour, then you decide — is the input that lets the framework work.
Fixr supports this through the validation-and-diagnosis flow: a vendor is dispatched for diagnosis, the actual repair cost is quoted after diagnosis, and the customer can decide to proceed, get a second opinion, or take the device for replacement. The diagnosis charge is small relative to the cost of making a wrong full-repair decision. The platform itself is free; the vendor's diagnosis and any subsequent repair are billed by the vendor directly.
Cross-link to asset register
This whole framework depends on knowing what's in your fleet, what each device cost, how old each is, and what repairs it has already had. A proper asset register makes the framework executable. See our GST-ready IT asset register template for the structure that supports this — and the AMS + Fixr integration post for closing the loop between asset records and repair tickets.
A 6-step decision flow you can use today
- Get the diagnosis quote (R). Don't make the decision on a guess.
- Pull the asset record: age, prior repairs, deployment role.
- Estimate L (remaining useful life) from age + prior issues.
- Estimate P (productivity loss) from employee role × likely future incidents.
- Compute N (new or refurb) net of residual value.
- Compare R + P vs N. Decide. Document the decision against the ticket for next time.
This sequence takes 15 minutes per decision and saves substantial rupees across a fleet of 50+ devices over a year.
FAQs
Do I need to do all this math for every repair? No. Below ₹2,000 quotes, just repair. Above ₹5,000, run the framework. In between, judgement call.
How do I estimate residual value? Search Cashify or OLX for the exact model in working condition. Take the median of three recent listings.
What about company-issued laptops being replaced for departing employees? Departures are a separate decision tree. The framework still helps if you're deciding whether to refurbish the returned device for the next hire or sell it.
Are refurbished devices reliable enough for office use? For non-frontline roles, yes — provided the refurbisher offers a 1-year warranty and uses business-class chassis (Latitude, ThinkPad, EliteBook), not consumer-class. Consumer-class refurbished is a worse bet.
How often should I refresh the whole fleet? A rolling refresh works better than a big-bang refresh: replace 20-25% of the fleet each year so the average device age stays around 2.5 years. This smooths the cash outlay and avoids cliff-edge failures.