Land Surveying Software: Crew Scheduling, Job Status and the Deliverable Backlog

Why job and crew software makes or breaks a land surveying firm

A surveying firm does not sell field time. It sells signed and sealed deliverables: an ALTA/NSPS survey, a topo with contours, a plat for recording, an as-built for a GC who is holding a pour. The crew hours are just the raw material. Which means the software that actually runs your business is not the data collector and it is not Civil 3D. It is whatever tracks the path from "client called" to "PLS sealed it and the invoice went out," and for most firms that path lives in a spreadsheet named something like Job Log 2024 FINAL v3.xlsx.

Here is the usual stack at a 40 to 120 person multi-office firm. Deltek Ajera or Vision for time and billing. QuickBooks if you are smaller. Trimble Business Center or Carlson or MicroSurvey for the data. Civil 3D or Carlson Survey for drafting. Dropbox or SharePoint for the shared drive. Smartsheet or a whiteboard for crew scheduling. Maybe Monday.com because someone tried it in 2022. The survey coordinator holds the whole thing together in their head and in a spreadsheet nobody else can safely touch.

The concrete scene: it is 3:40 on a Tuesday. Your coordinator is building Wednesday's crew board. Crew 3's total station went down at lunch, so Wednesday's ALTA in the next county needs a different instrument and a different party chief who is actually certified on the RTK network for that state. The client for that ALTA emailed a title commitment at 2pm with 31 Schedule B-II exceptions that somebody has to plot. Meanwhile a GC is calling about an as-built that was field-complete nine days ago and is sitting in the drafting queue behind four other jobs because nobody can see the drafting queue. Your coordinator spends 90 minutes on the phone and rebuilds the board twice. That is roughly 7 hours a week of a $95k person doing manual reconciliation, and it is not the expensive part. The expensive part is the two crew days lost to the mobilization that got resequenced badly, at $1,800 to $2,400 per crew-day of loaded cost, plus the ALTA that slips past the closing date and costs you the repeat client.

Problem: crew scheduling is a constraint puzzle, and your tools treat it as a calendar

Scheduling a survey crew is not scheduling an appointment. A crew assignment is only valid if seven things line up at once: the party chief holds the right license or certification for the state and job type, the crew has the right instrument (an RTK rover is useless in a heavy canopy job that needs a total station and a robotic setup), the drive time from the last job of the day is inside the DOT hours window, the site has right-of-entry executed, the utility locate ticket is active and not expired, the control network from a prior visit is still recoverable, and the client's access window is open.

ServiceTitan, Jobber, and Housecall Pro will let you drop a job on a tech's day. None of them know what a party chief is, none of them model instrument-to-job-type compatibility, and none of them know a locate ticket expires. Smartsheet is a grid, so it enforces nothing. Deltek schedules people against project budgets, not against equipment and site conditions. So the constraints live in your coordinator's head, which is why they cannot take a vacation and why the board collapses the day they are out.

What a custom build does: the crew, instrument, and certification are first-class entities with real attributes. Party chief record carries license number, state, expiration, and job types they are cleared for. Instrument record carries type, last calibration date, and current crew assignment. The job record carries required instrument class, required cert, ROE status, locate ticket number and expiry, and access window. Dispatch becomes a solver, not a drag-and-drop: it proposes tomorrow's board and shows the violated constraints in red before your coordinator commits. When Crew 3's total station goes down, the coordinator flags the instrument as out of service and the system immediately shows which four assignments across the next two weeks just became invalid, and proposes swaps that do not break drive time. This is the single highest-ROI thing we build in this category, and it is the thing no off-the-shelf tool will ever ship, because the constraint set only exists in surveying.

Problem: the job status everyone quotes is a lie

Ask five people the status of job 24-1187 and you get five answers. The PM says "field complete." The drafter says "waiting on the title commitment." Accounting says "70% billed." The client was told "end of week" by whoever answered the phone. The truth is the job has eight real states and your system tracks two.

Real survey job lifecycle: proposal sent, proposal accepted, research (deeds, plats, title work pulled), field ready (ROE and locates cleared), field in progress, field complete, data downloaded and reduced, drafting, QC review, PLS seal, delivered, invoiced. Deltek knows about the money. The shared drive knows about the files. Nobody owns the state machine. So the drafting queue is invisible and jobs die in it. We have seen firms where median field-complete to delivered was 14 days and nobody could name why, because there was no timestamp anywhere.

The build: one job record with an enforced state machine, each transition timestamped and attributed. Field crew closes out on a tablet and the job moves to "data download" automatically. The drafting queue becomes a real queue with a real backlog number and a real per-drafter WIP count. Now your coordinator can answer the GC's call in four seconds instead of four phone calls, and you can see, for the first time, that most of your cycle time is drafting and not field work, which changes where you hire next.

Problem: field data comes back as a pile of files with no chain of custody

Crew finishes, drives back or hits a hotel, and dumps a raw file, a set of field notes, 40 photos on the party chief's phone, and a sketch on a rite-in-the-rain pad. The raw file goes into Trimble Business Center. The photos go into a text thread. The sketch gets photographed or does not. Two weeks later during QC, the drafter needs to know whether that iron pin was found or set, and the note is ambiguous, and the party chief is 200 miles away on another job.

Dropbox and SharePoint are filing cabinets. They do not know that a photo belongs to a specific point number in a specific raw file on a specific job. Trimble Business Center knows the survey data cold but knows nothing about your job record, your client, or your invoice. That gap is where rework is born, and rework on a topo is a crew re-mobilization at real cost.

The build: a field app (native, offline-first, because your crews work where LTE does not) where every photo, note, and sketch is captured against a job and, where it matters, against a point number. Raw file ingest that parses the format your crews actually shoot in and links observations to the job record automatically. Monument records that persist across jobs, so when you go back to the same corner in three years, the prior recovery note and photo are already there. This is where a custom build compounds: your monument and control database becomes an asset that makes every future job in that county cheaper to field.

Problem: title commitments and legal descriptions eat senior hours

An ALTA arrives with a title commitment. Someone, usually a licensed or near-licensed person billing $110 to $160 an hour, reads Schedule B-II, identifies which of the 30-odd exceptions are plottable, pulls the referenced instrument numbers, and hand-keys the easement calls into the drawing. On a busy ALTA that is 3 to 6 hours before a single line gets drawn. Multiply by 20 ALTAs a month.

This is the one place in surveying where AI is unambiguously worth the money, and it is not a chatbot. Point a document extraction model at the title commitment PDF: it segments Schedule B-II, classifies each exception as plottable or non-plottable, pulls the recording instrument number and book/page, and extracts bearing-distance calls out of the referenced legal descriptions into structured data. It drafts an exception table with a confidence score per row. A human reviews and corrects; the corrections train the next pass. We are not claiming it drafts your survey. In our delivery experience it turns 4 hours of reading and transcription into about 40 minutes of reviewing a pre-filled table, and the transcription errors it removes are the errors that get you sued.

Two more places AI pays here, and only two. After-hours intake: a small commercial client calls at 7pm needing a boundary; a voice agent captures parcel ID, county, and job type, checks it against your service area, and drops a qualified lead on the board instead of a voicemail. And forecasting: with two years of your own job history, a model predicts field hours for a new job from acreage, county, canopy, and job type more accurately than a PM's gut, which is how you stop underbidding topos in wooded counties. Everything else marketed to you as AI in this space is autocomplete.

Problem: multi-office means multi-truth

Once you have three offices, each one has its own crew board, its own numbering convention, and its own idea of what "delivered" means. Corporate consolidates in Excel monthly. You cannot answer basic questions: which office has crew capacity next week, what is firm-wide drafting backlog, which clients are actually profitable after rework. Acquisitive firms have it worse, because the shop you bought last year is still on their old system and the integration keeps getting deferred.

Off-the-shelf multi-tenant tools force one workflow on every office, which is why rollouts fail: the Texas office genuinely does ROE differently than the Colorado office because state law differs. The build handles this with a shared core schema and per-office configuration: same job entity, same state machine, different required-field sets and different approval steps per office. One dispatch view with an office filter. One P&L that rolls up because the data model was designed to roll up, not because someone pivoted it.

What this actually costs and how long it takes

These are Digital Heroes bands from delivery across 2,000+ projects, not market averages.

A focused first release: crew and equipment scheduling with the constraint engine, the job state machine, offline field capture, and a client-facing status view. $60k to $130k, shipping in 12 to 16 weeks. That is a real production system for a firm running 6 to 15 crews, not a prototype.

A full platform: the above plus proposal-to-contract, project accounting integration or replacement, title document extraction, monument and control database, CAD deliverable pipeline hooks, client portal with deliverable download and e-signature, multi-office rollup. $150k to $400k, phased over 6 to 12 months. Phase it. Firms that try to ship all of it at once ship none of it.

What drives price up specifically in surveying:

Deltek integration. Ajera and Vision are the single biggest cost variable in this category. Their APIs are not friendly and the data model does not map cleanly to a job state machine. Budget 3 to 5 weeks for a real bidirectional sync, more if you are on an older on-prem Vision install.

Raw file format support. Every additional instrument ecosystem you support (Trimble, Leica, Topcon, Carlson) adds parsing and testing work. If your crews are mixed-fleet, say so on day one, not in week 10.

Offline-first field app. This roughly doubles field-app cost versus a web form, because conflict resolution when two crews sync the same job is real engineering. It is also non-negotiable if you work rural.

Multi-state compliance variation. ROE rules, locate ticket rules, and seal requirements differ by state. Each state you operate in adds configuration surface.

CAD integration depth. Reading job metadata into a Civil 3D or Carlson template is cheap. Pushing survey data into drawings is expensive and usually not worth it in phase one.

Build versus buy: take the position

Buy, genuinely, if: you run 5 crews or fewer, single office, mostly residential boundary and mortgage inspection work, one instrument ecosystem, one state. Your constraint set is small enough to live in a coordinator's head and a good spreadsheet. Deltek Ajera plus a disciplined shared drive plus a whiteboard will beat a $90k build, and you should spend the money on a second robotic total station instead. Anyone telling you otherwise is selling.

Build when you hit these signals, and you will know because you have hit them already:

Your coordinator cannot take a two-week vacation without the board degrading. That means the system is a person, and people leave.

You cannot state your drafting backlog in days without asking three people. That means your cycle time is unmanaged and you are losing repeat clients to slipped dates you did not see coming.

You are re-mobilizing crews more than twice a month for reasons that were knowable before the truck left: expired locate, missing ROE, wrong instrument. At $1,800 to $2,400 a crew-day, 24 avoidable re-mobs a year is $43k to $58k. The scheduling build pays for itself on that line alone.

You have acquired a firm, or you are about to. Integration is the forcing function, and bolting a second shop onto a spreadsheet does not work twice.

Your senior licensed staff are spending real hours on title commitment transcription. At 20 ALTAs a month and 4 hours each at $130 loaded, that is $124k a year of PLS-adjacent time doing data entry.

The honest version: most firms above 8 crews crossed this line two years ago and have been absorbing it in re-mobilizations and overtime ever since.

How to choose a developer for land surveying software

Ask them to draw the job entity for you before you sign anything. If their job record does not distinguish field-complete from data-reduced from drafting from sealed, they are going to build you a work order system with survey words painted on it. The state machine is the product. If they cannot articulate it in the first meeting, they will learn it on your budget.

Ask what they have integrated with Deltek. Specifically Ajera or Vision, specifically bidirectional, specifically in production. This is where these projects go over budget. A vendor who has never touched Deltek will quote you two weeks and spend six. Ask to speak to the client they did it for.

Ask how they handle offline conflict resolution. "The app caches locally" is not an answer. The answer you want describes what happens when a party chief edits a job on a tablet with no signal while the office edits the same job, and how the merge resolves without silently dropping a field note. If they have not thought about it, your crews will lose data in month two and never trust the app again.

Ask who owns the code, in writing, before the SOW is signed. You want the repository, the infrastructure accounts, and the right to hire anyone else to maintain it. A firm that keeps your survey job history and your control database inside their tenant holds leverage over you forever. Settle it in the SOW or find someone else.