Build vs. Buy: Should You Stand Up a Drone Team or Hire an Independent Pilot?

Good. Now we can talk.

---

The framing is wrong before the question gets asked

Most firms get this backwards. They start with the drone.

"It would be cool to have a drone."

That's the start of a six-figure mistake. I've watched it happen. More than once.

A drone is not a department. A drone is a sensor that produces data a department uses. The thing you actually need to build — if you build anything — is the decision workflow. The drone is downstream of that.

Same logic whether you're a small survey or engineering firm, an architecture studio, or an owner's rep juggling four jobs at once. The decision is the same:

Do you need a recurring capability in-house, or do you need a recurring outcome from someone who already built the capability?

That question matters more than the drone spec sheet.

---

You can't buy a drone and call it a VDC department

Here's the pattern. I've watched it happen — more than once this year alone.

Firm buys a Matrice 4 Enterprise. Maybe $5K for the basic combo, maybe $15K once you add a second Reach RS3, the six-battery kit, the GCPs, and the software seats. They hand it to the most junior person on staff, or — and this is the one that gets me — to the PE or PM who already has 200 things on their plate and a superintendent asking why the schedule is slipping. They fly once. They get a pretty ortho. They email it around.

Three months later, nobody's flying it. The drone's in a closet. The "drone program" is dead.

What went wrong?

• Nobody owned the deliverable. The ortho landed in someone's inbox. It didn't get tied to a change order, a progress claim, a sub dispute, or a haul-off volume. It was a picture. Pictures don't pay for themselves. Michael says this constantly and he's right — bottom line, it's always about the money.
• Nobody verified the accuracy. Pix4D says you're at a quarter inch. DJI Terra says the same. That's voodoo blackmagic until you put independent checkpoints on the ground and measure yourself. Most in-house programs skip this step. The data looks good until a sub disputes it, and then nobody can defend it.
• Nobody kept the cadence. Bi-weekly progress flights only work if they happen bi-weekly. One person out sick for a week kills the timeline comparison. The whole value is the time-series. Skip two cycles and you're starting over.
• Nobody stayed current on airspace. LAANC, NOTAMs, TFRs — they change. So do the FAA rules around them. The person who "flies on the side" doesn't have time to track that. The first time they hit a TFR they didn't know about, the drone's grounded and so is the program.
• Insurance. A consumer policy won't cover commercial operations. Most firms find this out the hard way — usually right after something goes wrong.

The drone didn't fail. The infrastructure around the drone failed. The drone was the easy part.

---

What it actually costs to build a real in-house program

When a firm tells Michael they're "going to start a drone program," here's what he asks. Run the numbers before you decide.

Hard costs, year one (every figure sourced, Sept 2026):

• Aircraft: DJI Matrice 4 Enterprise Basic Combo — $5,199 (Drone-Works). Bundle includes the M4E airframe, one battery, RC Plus 2 Enterprise controller, charger, and case. Cross-checked: DSLRPros lists M4E Universal Edition at $5,099 (aircraft only, currently sold out), GenPac lists the M4 series at $5,499–$7,849 pre-order. The M4E has built-in RTK, 4/3 CMOS 20MP mechanical-shutter wide camera, 48MP medium tele + 48MP tele, and laser rangefinder. It's the current enterprise mapping platform; the M3E is end-of-life.
• Batteries — minimum 3, working pros fly 6. The Basic Combo ships with one battery. Bi-weekly cadence on a real program burns through a single battery per flight with margin for wind, retakes, and weather holds. Three is the floor; six is what lets you finish a large site without a charging gap. M4E battery pricing wasn't extractable from the retailers I checked (DJI Store prices load via JS), but the comparable M3E battery kit is ~$660–$700 for three; M4E batteries trend higher per unit. Figure ~$200/battery as a working estimate, $1,000–$1,200 for the 6-battery kit a serious program needs.
• Emlid Reach RS3 — $2,999 each, you need two. Direct from emlid.com/reachrs3/, pre-tax. That's $5,998 for the base + rover pair. The RS3 is the current Emlid flagship — RTK/PPK GNSS receiver with centimeter-level accuracy on a 2-hour base-station cook. Worth noting: if you're only doing small sites where the M4E's built-in RTK is enough over CORS, you might get away with one RS3 + a CORS subscription. The two-unit setup is for when you need a local base to control the baseline yourself.
• Accessories (tablet, base station tripod, GCPs — lath, targets, paint — rover rod, pole, charging hub, hard cases): ~$1.5K–$2.5K. The GCP kit is the part most buyers under-buy. Targets, lath, paint, and a hard case so they survive the truck is a real line item.
• Software (Pix4D Mapper subscription ~$1.2K/year, DroneDeploy ~$99/month billed annually, Agisoft Metashape Pro perpetual at $3,499, or DJI Terra — free for the first year when bundled with the M4E purchase at DSLRPros; QGIS is free for verification): $1.2K–$3.5K/year.
• Commercial aviation insurance (liability + hull + payload, $1M policy): $1.5K–$2.8K/year — figure toward the higher end given the M4E hull value plus the Emlid pair plus the GCP kit.
• LAANC approvals, Part 107 prep, recurrent training, currency hours: $1K–$3K.

Realistic year-one total:

• Lean / minimum-viable (3 batteries, 1 RS3 if you use CORS, low-band everything): ~$14K
• Working professional setup (6 batteries, 2 RS3s, mid-band software and insurance): ~$20K
• Upper end (full accessories, Metashape Pro perpetual, top-band insurance, full training budget): ~$24K

For a serious bi-weekly mapping program, $18K–$22K is the honest working figure before a single flight.

Add LiDAR (SHARE S20 or equivalent), thermal (M4T instead of M4E — note the M4T chassis is a different line item entirely), or a Matrice 350 RTK bundle (~$16K, a different airframe class) and you're in a different bracket.

Soft costs, the part nobody budgets:

• One person's time. Real time, not "we'll fit it in." Budget 25–40% of an FTE for the first 12 months. Less than that and you don't get a program — you get a hobby that produces bad data.
• Recurring flight cadence on a real project means 2–4 hours per flight, every other week, weather-permitting, on top of the processing and the QA.
• A mentor or a peer group to keep the operator honest. (More on that in a second.)

If you're an architecture firm with two projects a year and zero in-house construction management, that math doesn't work. If you're a 40-person engineering firm with six active sites and a VDC team, it starts to make sense — but only if VDC owns the workflow.

---

So when does in-house actually win?

Three cases. Every one of them hinges on recurring volume and decision authority in-house.

1. You're a regional GC, ENR-ranked, with 8+ active sites and a VDC team that already owns model coordination.

The bi-weekly progress flight is one part of a larger workflow. The ortho feeds a model. The model feeds a clash detection. The clash detection feeds a coordination meeting. A drone pilot reporting to VDC — not the PM, not the marketing guy — makes sense.

If your VDC team is already running Revizto or Navisworks, and the drone output is one more input, an in-house operator is reasonable. You're not buying a drone program. You're extending an existing digital-twin program with a sensor.

You still need a Part 107 operator with the right temperament, insurance, and a checkpoint regime. You still need someone who'll catch a bad GCP day before the data ships to the PM. But you can find that person.

2. You're a survey or engineering firm where the ortho, the point cloud, and the linework are the deliverable — and the licensed surveyor signs off downstream.

Survey firms: I won't pretend to be one of you. But I'll say this — if your business model is already GIS-first and the drone is feeding your existing pipeline, you have the QA framework to make in-house work. You already verify accuracy. You already know what "tie to control" means. The drone is a new acquisition tool, not a new department.

Engineering firms with internal survey crews fall here too. The marginal cost of a drone is low because the verification muscle exists.

3. You need a full-time on-site presence for security, logistics, or compliance reasons.

Some sites — federal, secure, restricted airspace — don't lend themselves to a rotating independent contractor. If your security team needs to badge the operator, your ops team needs to schedule them, and your legal team needs them on your insurance, an employee is the right answer.

That's a rare case. But it's real.

---

So when does the independent win?

The rest of you. Most of you, frankly.

• The architecture studio with three projects a year.
• The small survey firm that doesn't have enough recurring flight work to justify a full-time operator.
• The owner's rep juggling four jobs at three different GCs, none of whom want a drone operator on their insurance.
• The engineering firm that needs bi-weekly flights for one specific site, not six.

For these buyers, building a team is a sunk cost that never earns back. You're paying $18K–$22K a year in hard costs alone to keep the capability warm for work that may or may not recur, plus the soft costs (an operator's time, training, currency, QA discipline). The independent carries the kit, the insurance, the currency, and the QA — and shows up on the cadence you need.

The math, bluntly: under ~5 missions a year, outsource without thinking. Between roughly 5 and 45 missions a year, fractional internal can pencil out — but you're still carrying the soft costs (insurance, training, the QA discipline) without the scale to amortize them. It's at ~45 missions a year and up that a dedicated full-time operator can beat vendor economics — if you keep them utilized and you're willing to own SOPs, compliance, and the processing pipeline. Below that, you're losing money on the program.

There's also a third option the question doesn't usually surface — and it's the one Wet Dog actually is for most of our clients. Long-term embedded independent. Not an employee. A contracted operator who knows your sites, your team, your cadence, your change-order process — but isn't on your payroll. The GC's flagship program in Denver has been on this model for years. It's the closest thing to in-house without the soft costs. If that's where you're leaning, ask about it specifically. Most independents will structure it.

---

What to ask before you decide either way

A short list. Run this whether you're leaning build or buy.

1. Who owns the data when it lands? If the answer is "we'll figure it out," pick the independent. If the answer is "our VDC team integrates it into the model by Tuesday," in-house is on the table.
2. Who's verifying the accuracy? Independent checkpoints. Not the software's own report. If nobody on your team knows what an independent checkpoint is, hire someone who does.
3. What's the recurrence? Under ~5 missions a year, outsource without thinking — the math won't pencil. Above ~45 missions a year with the same GCs and similar scopes, a dedicated operator (employee or long-term embedded independent) can beat vendor economics. Anywhere in between is where the decision actually lives — run the WDD-style TCO build with your real numbers before you commit.
4. Who's carrying the insurance and the LAANC approvals? If you have to add a drone operator to your existing policy, the underwriting question isn't trivial — most general-commercial policies weren't written with aviation exposure in mind, and you may need an aviation rider, an endorsement, or a separate policy. Ask your broker how long that takes and what it costs before you commit to the build. Annual drone-specific insurance for a small in-house M4E program typically runs $1.5K–$2.8K/year for combined liability + hull, given the aircraft value. Factor it.
5. What does "good enough" look like? If you're shipping an internal progress update the superintendent skims in the trailer, accuracy tolerance is wider than if you're feeding linework into Civil 3D. Decide before you buy the drone, not after.

If you can't answer those five, the answer isn't build or buy. The answer is talk to someone who already does this for a living — independent or in-house, I don't care which. The worst outcome is dropping $20K+ on an M4E rig because it felt like progress and then having it collect dust by spring.

---

The honest closer

I'll say it again, because it's worth saying once more.

If you're an architecture firm with one project a year — hire it out. If you're an owner's rep juggling four jobs at four different primes — hire it out. If you're a small survey firm dipping your toe in — hire it out for a year or two, learn what the data is worth, then decide whether to build.

If you're a regional GC with a real VDC pipeline and a flight cadence that's already scoped into the schedule — build, or embed a long-term operator. Either works.

If you're somewhere in between, run the five-question checklist. The math will tell you.

And if you do hire someone — independent, in-house, doesn't matter who — make sure they can answer #2 themselves: who verifies the accuracy? If they can't, find somebody else.

Bottom line money: the data is what matters. The drone is just how you get there.

---

Pro tip of the day: before you decide build or buy, count how many times this year you actually need a flight. Not how many times you "might want one." The number you write down is the number that decides the question.

---