Every few years a university asks whether it should build its own research information system instead of licensing one. It is a fair question, and the honest answer depends less on engineering capacity than on what the institution wants to own for the next decade.
Building is a real option, not a fantasy
Open-source CRIS platforms exist and are in production use. VIVO, a linked-open-data profile and research information platform, and DSpace-CRIS, an extension of DSpace maintained with 4Science and LYRASIS, both have documented institutional deployments — euroCRIS itself runs on DSpace-CRIS. Starting from one of these is a different proposition from writing a system from scratch, and it is the version of "build" most institutions should actually consider.
What you are really deciding
The licence fee is the visible part of the comparison, and usually the least interesting. The real question is which of these you want to own permanently:
- Data ingestion and matching. Pulling records from Scopus, OpenAlex, ORCID, Crossref or Web of Science is straightforward once. Keeping it correct is not: author disambiguation, deduplication across sources, affiliation changes and retractions never stop arriving.
- Schema drift. Every external source changes its API and its fields. Someone has to notice, and fix it, before the annual report is due.
- Reporting demands. Accreditation bodies and ranking organisations change what they ask for. A built system must follow those changes on their schedule, not yours.
- Institutional memory. The developer who understands the matching rules will eventually move on. Documentation written for that moment is the difference between a system and a liability.
Where build genuinely wins
- You have unusual requirements no vendor serves. Local languages, national reporting formats, or a research culture where standard fields do not fit.
- You have durable in-house capacity. Not one enthusiastic developer — a team, with a budget line that survives leadership changes.
- Data sovereignty is absolute. Some institutions genuinely cannot send metadata off-premise, although several commercial systems now support on-premise deployment too.
- You already run an institutional repository well. If DSpace is healthy and staffed, DSpace-CRIS is an incremental step rather than a new discipline.
Where build usually fails
- The pilot succeeds and the maintenance does not. The first version is the cheap part. Year three, with nobody assigned, is where systems quietly die.
- Data quality is treated as a technical problem. It is mostly a governance problem: who decides the authoritative publication count, and who fixes a wrong record.
- The cost comparison ignores staff time. A developer-year is not free simply because it is already in the payroll budget.
- Reporting deadlines arrive faster than fixes. Accreditation does not pause while a bug is triaged.
A fair way to compare
Compare over five years, and include the lines institutions routinely leave out:
- Licence or subscription, per year
- Implementation, migration and integration — one-off but substantial, and charged per feed or per data type by most vendors
- Internal staff time, for both options, costed honestly
- Underlying data subscriptions, which may already sit in the library budget
- The cost of a year without the system, if a build slips
Our guide to what a research information system costs collects the public contract values and published rate cards, which gives the buy side of that comparison real numbers rather than guesses.
The middle path most institutions miss
Buy the ingestion, matching and reporting engine; build the parts that are genuinely local. A national reporting export, an internal dashboard for deans, or an integration with a home-grown HR system are all reasonable things to build on top of a platform you did not write.
The test is simple: build what encodes your institution's own rules, and buy what every research university needs to do identically.
Frequently asked questions
Is open-source CRIS software free?
The licence is free; the system is not. VIVO and DSpace-CRIS both require hosting, configuration, data ingestion work and ongoing maintenance. Treat them as a different cost structure rather than an absence of cost.
How long does building take?
There is no credible generic answer, and any vendor or consultant offering one is guessing. What can be said is that the first working version is a small fraction of the total effort: matching rules, deduplication and reporting formats consume far more time than the initial data pipeline.
Can we start by building and move to a product later?
Yes, and it is a reasonable hedge if you keep your data clean and exportable. Insist on documented schemas and regular exports from day one, whichever direction you eventually take.
What does Discover RIMS do in this decision?
It is the buy option, with one difference worth weighing: it can run on-premise or hybrid, so data sovereignty alone does not force a build. If you are comparing us against building, ask for the list of matching rules and export formats — those are what you would otherwise be writing yourself.
Related reading
- Do you need a RIMS? A readiness assessment
- RFP evaluation criteria for a RIMS
- How Discover RIMS compares with Pure, Converis and Esploro
Sources
- VIVO project — vivoweb.org
- DSpace-CRIS documentation — wiki.lyrasis.org
- euroCRIS, which runs on DSpace-CRIS — dspacecris.eurocris.org
Checked September 2026.