How to Choose an ELN: 5 Checks Before You Buy
How to choose an ELN is less a question about features than about evidence: five checks decide whether an electronic lab notebook will survive a Korean national R&D verification — trustworthiness mechanics, guideline fit, the confirmer signature workflow, pricing structure, and whether your data can leave when the contract ends.
The reason a checklist beats a feature matrix is that every product looks identical in a demo. Each one shows a clean notebook page, a file drag-and-drop, a search bar. The differences surface months later, in the specific moments the notebook exists for — a ministry inspection, a misconduct allegation, a patent dispute — and by then you have a year of records committed to whichever system you picked. The five checks below are the ones that predict how a system behaves in those moments. Run them before you sign, not after.
How Do You Choose an ELN? Five Checks Before You Sign
1. Does it meet the trustworthiness requirements?
An ELN is not a note-taking app with extra steps; it is a machine for proving three things about every entry: when it was written, who wrote it, and what has happened to it since. Ask the vendor to demonstrate each one. Timestamps should be set by the server, never editable by the author’s clock. Every entry and every edit should attach to a named individual on an account that is never shared. And corrections should preserve the prior version — a correction you cannot see is an overwrite, and an overwrite is exactly what a verification is designed to catch. If the answer to any of the three is “the lab manager handles that,” the system does not do it.
2. Does it map to the national R&D guideline?
The Guideline on Lab Notebooks for National R&D Programs(MSIT Notice No. 2021-102) does not name tools — it names properties, and leaves each institution to guarantee them in its own regulation. The head of the institution must be able to confirm a record’s date and recorder (Art. 7(2)); the guideline defines separate recorder and confirmer roles (Art. 2); the default retention period is thirty years (Art. 10(3)); and the responsible ministry can inspect whether the system actually operates, not merely whether the regulation exists (Art. 13). A product that was not built around these provisions will leave your institution’s regulation promising things the software cannot deliver. For the clause-level breakdown of what the guideline requires, see electronic lab notebook requirements in Korea.
3. Can a confirmer actually sign — inside the tool?
The guideline’s two-role structure (Art. 2) means the notebook is not complete when the researcher hits save; it is complete when someone else reviews and signs. So the selection question is not “does the ELN support signatures” but whether the whole confirmation loop lives in the system: can the principal investigator see which entries are pending, sign under their own identity, and reject one with a reason that stays attached to the record? Watch what happens on paper in most labs — entries get printed, signed by hand, and filed separately — and you are looking at a system whose real notebook has quietly moved outside the software. If confirmation happens outside the tool, the tool is not the notebook; the printouts are, and you paid for a wiki.
4. What does the price actually scale on?
ELN pricing comes in three shapes: per seat, per project, and storage-based — and the cheapest-looking shape depends entirely on your lab’s biology. A per-seat price punishes the lab that rotates through master’s students and postdocs every two years; a storage price punishes the lab that attaches microscope images to everything; a per-project price punishes the lab with fifteen small industry contracts. Ask for a three-year cost at your realnumbers — current headcount including incoming students, actual monthly data volume, project count — not the brochure seat price. Then add the costs nobody brochures: onboarding, template setup, and what an export or migration costs if you ever leave. “Free” tiers deserve the same question; free until the storage cap, or free until the feature you actually need, is a price you have not been quoted yet.
5. Can your data leave — and survive for thirty years?
The default retention period of thirty years (Art. 10(3)) will outlast several vendor contracts, possibly the vendor. So the exit question is not pessimism; it is the retention requirement stated plainly. Ask to see an export: what format comes out, is it readable without the software, does the export preserve timestamps, authors, signatures, and the revision history — or does it flatten your notebook into a pile of PDFs that prove nothing? A system whose records are only legible inside itself is not preserving your data; it is holding it. Whatever you buy, buy with the exit in view, and put the export path in the contract.
Korea ELN Comparison: Three Species of Product
The Korean market splits into three types, and the five checks above hit each differently.
Domestic cloud services— such as 구노 (GOONO) — are built around the Korean national R&D submission context: cloud-based, with the three statutory elements (time certification, author signature, third-party signature) handled by the product itself, and file upload streamlined for report assembly. Their strength is that the compliance target is local; their limit is usually depth — check what the audit trail and export look like under load, not just the submission flow.
Open-source ELNs— eLabFTW being the most widely used — are free and self-hosted, which makes them the natural home of the “why pay a subscription” argument. The honest accounting: you inherit the server, the backups, the access control, and the burden of proving the audit trail’s integrity to an inspector yourself. Free software, paid operations. For a lab with no IT staff, that trade is often worse than it first looks.
Global commercial ELNs— Benchling (biology-focused, with free academic plans), Signals Notebook, LabArchives — bring enterprise maturity and deep domain tooling. What varies is how they map to the Korean guideline’s specific requirements, and what support looks like in Korean, at your timezone, when the inspection is next month. Run check #2 and #3 against them explicitly rather than assuming enterprise means compliant here.
None of these types is universally right; the five checks are how you decide which trade-offs your specific lab can live with.
FAQ
What does an ELN cost?
It depends on the pricing shape — per seat, per project, or storage-based — and on your lab’s size, turnover, and data volume more than on the vendor’s list price. The only number worth comparing is a three-year total at your real headcount and data volume, including onboarding and any export or migration fees. Any quote that is not that number is not yet a price.
Is a free or open-source ELN enough?
Sometimes, yes — a technically strong lab with real IT support can run eLabFTW well. But “enough” has to include the operational work: hosting, backups, access control, and demonstrating the record’s integrity to a verifier. Budget that labor honestly before concluding the software being free makes the system free.
What should we ask for in the demo?
Three things vendors rarely volunteer: show me a correction (does the prior version survive?), show me an export (does it prove anything without your software?), and show me a confirmer rejecting an entry (does the review loop live in the tool?). The demo that only shows a clean page being written is showing you the part every product can do.
If you would rather evaluate a system where attribution, server-side timestamps, and a tamper-evident history are the defaults the demo has to live up to, pricing has the plan breakdown.