In a typical first-round slate of eight, I see two or three candidates open a software or device case with a retail profitability tree. They talk "price, volume, cost" as if the client sold soap. The sheet is listening for something else: whether you can read an ARR bridge, separate net revenue retention from logo count, and price cloud COGS that move with usage.
That is the thesis of this guide. A technology case interview is a unit-economics problem dressed as strategy. Software, platforms, and hardware-plus-subscription products live on recurring contracts and attach rates. Telco ARPU, streaming ads, and cable bundles belong in a TMT case. A bank swapping its core system belongs in digital transformation. After this piece you should be able to structure a SaaS or device case the way an interviewer scores it, reproduce one original mini-case with consistent invented numbers, and know which public filing metrics set a realistic bar.
What the interviewer is actually scoring
The prompt usually sounds strategic: "Should we discount to win logos?" "Seat vs usage?" "Cut device ASP to drive software attach?" Your answer gets points when you convert that prompt into a renewable-revenue model before you pitch a narrative.
I listen for five moves, in roughly this order:
- Name the revenue stock (ARR or product revenue run-rate) and how it moves (new, expansion, churn, downsell).
- Separate logo retention from dollar retention.
- State the packaging metric (seats, usage, platform fee, or device attach).
- Put a number on sales efficiency (CAC payback or a rough magic number).
- Treat gross margin as variable with hosting, support, and—if relevant—inference, not as "software is 80% forever."
Miss the first two and the rest of the case is theater. Hit them early and the interviewer often hands you the next exhibit.
Primary benchmarks, not folklore
Public SaaS filings give you the range candidates invent when they guess. Use them as anchors, then invent original case numbers around them.
| Company / filing | Metric | Figure (as of date in source) |
|---|---|---|
| Snowflake FY2025 results | Net revenue retention | 126% as of Jan 31, 2025 |
| Snowflake same release | GAAP product gross margin | 71% for FY2025 |
| Salesforce FY2025 earnings exhibit | Subscription and support GM | ~82.6% ($35.7B revenue / $6.2B cost) |
| Salesforce same exhibit | Remaining performance obligation | $63.4B as of Jan 31, 2025 |
| Datadog S-1/A | Dollar-based net retention | 146% as of Jun 30, 2018 and 2019 |
| Datadog FY2024 results | GAAP gross margin | 81% for 2024 |
| Apple FY2025 10-K product mix | Services vs iPhone | Services $109.2B; iPhone $209.6B |
Apple does not publish a clean "services attach rate per iPhone." What the filing does show is that services are a large, growing companion to device sales—the hardware case pattern is BOM and ASP on the device, plus a recurring attach you have to model yourself.
Ned's rule. If you cannot write beginning ARR, ending ARR, and the four bridge lines between them on one napkin, you are not ready for the recommendation. Strategy slides without that bridge are decoration.
The units that separate pass from fail
ARR vs bookings vs billings. Bookings spike when a three-year prepaid deal closes. ARR is the annualized run-rate of recurring revenue. Mixing them is how candidates invent fake growth.
NRR vs GRR. Net revenue retention includes expansion. Gross revenue retention does not. A 120% NRR with an 85% GRR is a company being rescued by a few expanding accounts while the mid-market leaks.
Seats vs usage. Seat cuts can raise consumption if the value metric is usage. Usage packaging can also destroy margin if inference or egress sits in COGS and you never asked for the unit cost.
Hardware ASP vs attach. Device revenue is ASP × units. Software economics ride attach rate × ARPU per active device. Cutting ASP to "drive volume" without modeling attach and SaaS margin is how you shrink total gross profit while celebrating units.
Worked mini-case: cut the gateway, grow the attach?
Prompt. OrbitGate sells industrial IoT gateways plus a monitoring SaaS plan. Invented but consistent figures for one fiscal year:
| Line | Value |
|---|---|
| Gateways sold | 220,000 |
| Hardware ASP | $260 |
| Bill of materials (BOM) | $92 per unit |
| Paid SaaS attach | 58% of devices sold this year |
| SaaS price | $72 per device per year |
| SaaS cloud + support COGS | 22% of SaaS revenue |
| Sales proposal | Cut ASP 15% to $221; expect volume +20% to 264,000; attach rises to 74% |
Sales argues hardware is a loss-leader for software. Do you take the cut? Assume one-year SaaS contracts on newly activated devices, no MFN on the installed SaaS base for this year, and no change in BOM.
Step 1 — Current economics
Hardware revenue: 220,000 × $260 = $57.2m Hardware COGS: 220,000 × $92 = $20.2m Hardware gross profit: $37.0m
SaaS revenue (new-year activations): 220,000 × 0.58 × $72 = $9.19m SaaS COGS at 22%: $2.02m SaaS gross profit: $7.17m
Total gross profit: $44.2m Blended hardware + SaaS gross margin on $66.4m revenue: 66.5%
Step 2 — Proposed economics
Hardware revenue: 264,000 × $221 = $58.3m Hardware COGS: 264,000 × $92 = $24.3m Hardware gross profit: $34.0m (down $3.0m)
SaaS revenue: 264,000 × 0.74 × $72 = $14.07m SaaS COGS: $3.10m SaaS gross profit: $10.97m (up $3.80m)
Total gross profit: $45.0m — about +$0.8m versus today.
Step 3 — What the sheet rewards next
The cut barely pays in year one. The recommendation is not a heroic yes. It is conditional:
- Accept only if Sales can prove attach actually reaches ~74% in a pilot segment (not a slide assumption), and if SaaS COGS stay near 22% when usage scales.
- Require a floor ASP for enterprise SKUs so BOM-plus-channel cost does not go underwater if attach slips to the mid-60s.
- Ask for the second-year SaaS renewal on the enlarged installed base; that is where the proposal earns or dies. If year-two NRR on SaaS is only 100% with 10% logo churn, you bought units and still have a thin software franchise.
- Reject an across-the-board list cut if the mid-market already attaches above 70%; then you are giving away hardware margin for attach you already have.
A generic "grow volume" tree would have celebrated +44,000 gateways and missed that hardware GP fell $3m while software had to carry the case.
Packaging, sales efficiency, and cloud COGS
When the client is pure SaaS, open with packaging and sales efficiency before "market share."
Seat vs usage. Seat pricing is easy to forecast and easy to discount into the installed base. Usage pricing aligns with value and with cloud COGS. In the room, force the interviewer to tell you which metric customers already buy on. Then ask whether a price experiment on new logos will be most-favored-nationed to renewals. Enterprise sales leaks; assume it does unless told otherwise.
CAC payback. A rough case version: sales and marketing spend last year divided by net new ARR, adjusted for gross margin if you have time. A payback that stretches past roughly two years on a mid-market SaaS motion is a cash incinerator dressed as growth. Public companies publish different definitions; in a case, define yours once and stay consistent.
Gross margin. Mature multi-product SaaS can sit near the low-to-mid 80s on subscription, as Salesforce's FY2025 subscription-and-support lines imply. Usage-heavy data platforms often print lower product margins—Snowflake's FY2025 GAAP product gross margin was 71%. If the case adds an AI feature, ask for inference cost per query before you bless an 80%+ story.
What a CPG profitability tree misses
- Volume is not logos if price on the base falls. ARR is price × quantity × expansion, with churn as a negative.
- Replacing a lost logo reloads CAC; churn is a revenue and a sales-efficiency problem.
- Multi-year bookings can mask rotting NRR for two renewals.
- Platform and marketplace effects (developers, ISVs) do not appear in a three-box cost tree.
- Hardware attach is a second P&L, not a footnote under "other revenue."
If you treat OrbitGate like a supermarket, you cut sticker price, celebrate units, and never check whether SaaS gross profit covered the hardware giveaway.
Practice this on a live case
Run one scored case where you are forced to build an ARR or attach bridge before the recommendation. Start with the five-minute typed rep at /start, or take the ~18-minute voice case at /interview. Everything is open for seven days, no card; then $120 for a recruiting season or $49 a month. For math reps on unit economics without a full case, use /drills/math.
Frequently asked questions
What is a technology case interview in consulting?
It is a case about a software, platform, or hardware-plus-subscription business. Interviewers expect recurring-revenue math—ARR bridges, net revenue retention, packaging, and cloud or device COGS—not a media ARPU case and not a corporate transformation roadmap.
How is a SaaS case different from a TMT case?
SaaS cases turn on seats or usage, NRR, and sales efficiency. TMT cases (telco, media, ads) turn on ARPU, churn of consumers, content or spectrum costs, and ad yield. If the client is a mobile operator or a streamer, use the TMT frame instead.
What metrics should I memorize for tech cases?
Know how to compute ARR bridges, NRR vs GRR, a simple CAC payback, and gross margin with variable cloud COGS. Use public filing ranges (for example Snowflake's 126% NRR and ~71% product GM in FY2025) as sanity checks, then invent case-specific numbers the interviewer gives you.
Do hardware product cases belong in technology interviews?
Yes, when the economics are device BOM, ASP, and software or services attach. That is still product unit economics. It is not a telecom network case and not a "digitize the plant" transformation case.
Should I open with a profitability framework?
Open with the revenue stock and how it moves, then margin and sales efficiency. A generic profitability tree is fine as a checklist after those lines exist. Leading with it usually misses NRR and attach.
Sources
- Snowflake Reports Financial Results for Q4 and FY2025 — net revenue retention 126% as of Jan 31, 2025; product revenue and RPO context. Checked 2026-09-24.
- Snowflake FY2025 Q4 earnings exhibit (SEC) — GAAP product gross margin 71% for FY2025. Checked 2026-09-24.
- Salesforce Q4 and FY2025 results exhibit 99.1 — subscription and support revenue and cost; RPO $63.4B. Checked 2026-09-24.
- Datadog S-1/A (Sep 2019) — dollar-based net retention 146% as of Jun 30, 2018 and 2019. Checked 2026-09-24.
- Datadog FY2024 financial results exhibit — GAAP gross margin 81% for 2024. Checked 2026-09-24.
- Apple Inc. Form 10-K (fiscal 2025) revenue disaggregation — Services and iPhone net sales for hardware-plus-services framing. Checked 2026-09-24.
CoachNed is independent and unaffiliated with McKinsey, BCG, Bain, or any other firm named for interview-style practice.