Geotechnical Data Management Software: Keep gINT, or Build the Archive Your Firm Actually Needs?

When is off the shelf genuinely the right call here?

If you run under about 15 ground investigations a year, use one or two laboratories and produce logs to a single common standard, buy. Bentley gINT remains widely used, OpenGround is the current path in that ecosystem, HoleBASE SI is strong in markets built on the AGS data exchange format, and Datgel builds serious extensions on top of both. The licence cost is modest against any build, and there is a pool of engineers who already know these tools, which matters every time you hire.

Log drafting in particular is not something to rebuild. A borehole log is a drawing governed by hatching conventions, column order, how a water strike is shown, whether standard penetration test results appear as increments or a single value, and which laboratory results print alongside the description. Those products carry years of accumulated detail in exactly that area, and any developer who proposes to reproduce it inside a first release has not understood what they are quoting.

Buy also if your firm has no appetite to own a system. A data platform with nobody responsible for the taxonomy, the presentation standards and the laboratory profiles degrades within two years regardless of how well it was built. A packaged product with a vendor maintaining it is the honest answer for a firm that will not fund an internal owner, and pretending otherwise wastes a six figure budget.

The test that settles it: can an engineer starting a new job find your firm's prior investigations near that site inside ten minutes. While the answer is yes, or while you rarely work the same ground twice, you have not outgrown what you can buy.

When does a custom build actually pay off?

The signals are commercial and they usually show up on a timesheet before anyone frames them as a software problem.

Laboratory data entry has become a recognisable cost line, meaning a graduate engineer spends real weeks a year retyping results. You receive results from more than three laboratories in incompatible formats, one returning a workbook, another a report with a spreadsheet attached whose columns are in a different order. You hold a decade or more of investigations in a region you keep working and cannot search any of it. A major client requires deliverables your current tool cannot produce without manual editing. Or a professional liability review has made undocumented parameter selection uncomfortable.

The cause is that packaged tools are built around the current project. They capture data, produce a log and export a section, and they do that well. What they are not built for is a firm treating its investigation history as an asset, absorbing laboratory feeds that will never standardise, or producing the specific deliverable one long standing client has demanded for fifteen years.

The archive is the sharpest version of this. An engineer nearly certain there are boreholes within 300 metres from a job in 2011 spends forty minutes searching a network drive, gives up, and specifies four more holes. The client pays for ground investigation the firm already did. That happens quietly, repeatedly, and it never appears as a line item anywhere. Twenty years of investigations across a city is a genuine advantage: it lets you propose a smaller, smarter scheme and de risk early advice. It is worth nothing while it is unsearchable.

How do they compare on the things that matter in this industry?

• Log presentation. Packaged tools handle standards through templates and do it reasonably well. Where teams lose days is the last mile, a client whose specification needs a column the template cannot express, so the log gets exported and then edited. Once it is edited it has diverged from the data, and one of the two versions gets issued.
• Laboratory feeds. The AGS format solves this when everyone uses it, and adoption varies by market and by laboratory. A small specialist laboratory sends a report because that is what their instrument software produces, and asking them to change output for one client is a conversation that ends politely and changes nothing. An ingestion profile per laboratory costs $6,000 to $10,000 and is the only reliable answer.
• Sample identity. Laboratories return results under their own numbering and clients renumber holes between phases. Identification has to be composite, meaning project, hole, depth range and sample reference. A developer proposing a single system issued identifier has not dealt with a real laboratory return.
• Historical search. No logging package indexes your archive spatially, because the archive is a firm level asset and the packages are project level tools. This is a category gap rather than a feature gap.
• Parameter provenance. Characteristic values come from test results filtered by judgement, and that judgement usually lives in a spreadsheet on one engineer's machine. No packaged tool attempts to record it, which is exactly why it carries professional liability.
• Exit path. Ask your incumbent whether extracting your own data back out is included or chargeable, and in what format. That answer is part of the licence price whether or not it appears on the invoice.

What does total cost of ownership look like at your scale?

From Digital Heroes delivery experience, a first release covering the project, hole, sample and test data model, tablet field logging with genuine offline capture, laboratory ingestion profiles with exception reporting, and automated log and section production in one or two of your standard presentations runs $60,000 to $140,000 over 12 to 16 weeks. Adding a spatial historical archive with extracted data, several client specific deliverable formats, laboratory scheduling with turnaround tracking and export into design software takes it to $160,000 to $280,000 over 6 to 9 months. A large scanned archive, instrument level integration and parameter selection with recorded justification pushes it to $280,000 to $420,000 over 9 to 12 months.

A worked shape: a regional consultancy running about 60 investigations a year, four laboratories, three client presentation standards and twenty years of digital reports concentrated in two metropolitan areas came in at $188,000. Drop the archive and the map search and the same firm is at $136,000, which is a defensible place to stop for a year.

On the running side, budget 15 to 20 percent of build cost a year, and in this category most of it is laboratory maintenance. Laboratories change export formats, change instrument software or start reporting a new test, and each of those breaks an ingestion profile until somebody fixes it. Field tablets have a genuinely short working life in ground investigation and belong in the same business case even though they are not software.

The comparison is less clean here than in most sectors, because your gINT or OpenGround seats stay on the renewal either way. Do the arithmetic on labour instead. For fifteen investigations a year the packaged licence wins over three years and it is not close. For sixty, with four laboratories and three presentation standards, three years of licence plus seat growth plus presentation change requests against a build plus three years of support usually looks worse in year one and better by year three, with the archive value sitting entirely outside that sum.

What does the hybrid look like, and when is it the honest answer?

For most consultancies in the middle of this range the hybrid is the recommendation. Keep the logging package. Build the layer it was never designed to be.

The split is clean. gINT or OpenGround keeps owning log and section drafting, which is where the accumulated detail lives and where your engineers already know what they are doing. The custom layer owns laboratory ingestion with exception reporting, the composite sample identity model, the spatial archive with search, and parameter selection with recorded justification. Validated data flows into the drafting package as an output channel rather than being retyped into it.

There is a smaller and often better hybrid worth naming, and it is the highest return spend in this category. Build the archive extraction and spatial search on its own, without touching your logging tool at all. Extract location, depth, stratum descriptions and headline test results, index them spatially, and let a new site boundary return every prior investigation within a radius with a link to the original report. That is a $30,000 to $45,000 workstream for two metropolitan areas of digital reports plus a modest search interface, and it changes how you scope and price work.

Two disciplines make the hybrid hold. Order archive extraction by value, meaning the corridors and cities where you bid most often, not chronologically, and stop when the next region no longer pays for itself. And accept partial extraction. You do not need every value, because the engineer will open the source report anyway. Chasing completeness is precisely where these projects overrun by six figures.

Which should you choose, by operator size and stage?

Under 15 investigations a year, one or two laboratories, one presentation standard: buy. gINT, OpenGround, HoleBASE SI or a Datgel extension will serve you properly and nothing else here applies yet.

Any firm without an internal owner for the system: buy, whatever your volume. This is the failure mode that wastes the most money in this category, and it is entirely predictable in advance.

Roughly 15 to 40 investigations a year, two or three laboratories: buy, then measure one thing. Count the hours a year your graduates spend retyping laboratory results and reconciling sample references. If that is under a couple of weeks in total, keep configuring the tool you own.

Above 40 investigations a year with three or more laboratories: build the ingestion and data model layer, keep the drafting package, and support your dominant client's presentation standard properly before adding the rest.

Established firms working repeatedly in one geography, at any volume: consider the archive project on its own merits, separately from any platform decision. If the archive is the reason you are reading this, a focused extraction and indexing project is likely the right answer and it costs a fraction of a platform.

Firms carrying a professional liability concern about design parameters: build the parameter selection workflow, and put it in phase one rather than phase three. Recording which results were included, which were discarded and why, with a permanent link to the exact tests behind a characteristic value, is the piece no packaged tool attempts and the piece a technical director will sponsor once they have sat through one review.

If you want a second opinion before signing anything, Digital Heroes builds and runs its own products, so the people choosing your architecture live with those decisions on their own revenue. The document is yours whichever way you go.