1. Digitisation Doctor
Wellcome Trust, 13 April 2013
Ed Fay, Digital Library Manager
e.fay@lse.ac.uk | @digitalfay
Digitisation: Costs and
Sustainability
2. Costs of Digitisation
what will it cost you?
nature of the originals
types of content
purpose and user needs
…it depends!
3. Costs of Digitisation
Management
planning, fundraising, procurement, user studies, copyright
Content creation
handling, preparation, scanning, metadata, workflows
Access provision
user studies, system maintenance
Preservation
storage replacement, format management
5. Who funds what?
Be explicit.
“Many projects draw some form
of host support, but it is not
budgeted or guaranteed.”
(Ithaka, 2013)
6. Who funds what?
Grants
Contributions from partner organisations
Core budget from your institution
In-kind support
Earned income (revenue from sponsorship,
advertising, subscriptions)
Donations or individual philanthropy
(Ithaka, 2013)
7. Short-term vs long-term
What happens after
the project?
Budget spent.
Skilled staff leave.
No money, thousands of images.
9. Sustainability
Project management
Marketing and outreach
Content selection
Legal services
Content production (eg scanning, metadata creation)
Financial and accounting services
Information technology and support services (eg server
space, tech support, upgrades, etc)
Indirect costs (eg office space, supplies)
Digital preservation
Other (please describe)
12. In-house vs out-source
Reasons for in-house:
Skills retention.
Few format types.
Long timescales.
Very small or high volume.
Research and development.
Can fund staff and/or equipment.
13. In-house vs out-source
Things to think about:
Equipment purchase and maintenance.
Skills development (training).
Diversity of content.
Capacity and throughput.
Future demand (or lack of it).
14. In-house vs out-source
Reasons for out-source:
Skills recruitment.
Variety of format types.
Short or long timescales.
Low or high volume.
Known specification.
Can fund capital spend.
15. In-house vs out-source
Things to think about:
In-house support.
Spec writing.
Tendering or procurement.
Quality criteria.
Recurrence.
Diversity of format types.
16. In-house vs out-source
Last few years at LSE:
19thC and 20thC bound volumes. 19thC manuscripts (loose
leaf and bound). 20thC newspapers. Glass plate negatives.
Nitrate/acetate/35mm negatives. Photographic prints and
albums. Large format maps. Posters. Pamphlets. Badges and
postcards. 3D objects (plates, purses, boxes).
Worked with around a dozen different suppliers,
each bringing the right equipment and experience.
20. Short-term Long-term
User studies Requirements gathering On-going engagement
Copyright
None
None
Rights clearance
Metadata Description Technical/Preservation
Image capture Scanning (+ conservation?) None (rescan)
Workflow
Automated (e.g. batch OCR)
None
(reprocess)
Manual (e.g. transcription)
Storage
Initial capacity (incl backups)
Maintenance/cloud access
Capacity increase
Replacement cycle
Online access Commercial or open-source
UX updates
System replacement
Digital preservation Planning (e.g. standards) Actions (interventions) ???
Costs of digitisation
Low Medium High
21. Cost models
17 different cost models for digital preservation
http://wiki.opf-labs.org/display/CDP/Home
“We worked hard to tackle this particular
challenge in LIFE3, and based on the
evaluation project that followed, we almost
completely failed.”
(Wheatley, 2012)
22. Costs of Digitisation
The golden rule for cost
reduction is to look at every
stage that requires human
intervention and either remove
it, reduce it or make it as
efficient as possible.
(Minerva, 2006)
23. Costs of Digitisation
Methods to reduce cost:
1) Reduce the cost of labour;
2) Automate to reduce levels of human intervention in digital
conversion and metadata creation;
3) Select and prepare originals to enable higher volumes and
reduce variation in the workflow;
4) Increase overall performance and throughput to make the
most efficient use of capital expenditure; and
5) Continuous improvement and optimisation through rigorous
quality assurance.
(Minerva, 2006)
24. “Stewardship is easy and
inexpensive to claim; it is
expensive and difficult to
honor, and perhaps it will
prove to be all too easy to
later abdicate.”
(Lynch, 2003)
25. Cost everything.
Even if you don’t have to cash fund it now, you
may have to make the case in the future.
Plan for the long-term.
Think about what happens after the project.
At the extreme, this is about repositioning
the organisation to make digital part of your
business-as-usual.
Final thoughts
26. References
Sustaining our Digital Future: Institutional Strategies for Digital Content
(Ithaka, 2013)
Handbook on Cost Reduction in Digitisation
(Minerva, 2006)
Rapid Capture: Faster Throughput in Digitization of Special Collections
(OCLC, 2011)
Toolkit for the Impact of Digitised Scholarly Resources (TIDSR)
(Oxford Internet Institute, 2011)
Balanced Value Impact Model
(Tanner, Kings College Digital, 2013)
27. Ed Fay, Digital Library Manager
e.fay@lse.ac.uk | @digitalfay
?
Editor's Notes
Return to Simon’s introduction: IT DEPENDS- what you have, what you are trying to do with itCosts associated with a whole host of activities, well beyond “item on scanner”Management—some one-off, some on-going part of programme of workContent creation—all detail of workflows we’ve heard about todayOnce you have digital content—whole set of other problems (how to provide access, how to sustain that access over time)Approach costs from 3 angles…Digitisation is likely to be a small, perhaps growing part, of any organisation that already funds a wide variety of activities.Digital is emerging as an extension to these activities, so you must be clear about what must be funded that is different, and where you can rely on existing resources.(This might mean people’s time. It might mean IT infrastructure.)But the important point–know what is involved in every aspect of supporting digitisation, and make sure you understand how it is funded.If someone asks the question in the future (why is my staff member working on this? or why is our storage full of all these TIFF files?) the answer “I assumed it was taken care of” may not get the response you are looking for.Of course, this is about pragmatism as well. Presumably it is unlikely that you will be asked to takes things like building or estates costs into account (although it can and does happen).But what about electricity to run servers? Network bandwidth? IT infrastructure? Scanning equipment? Etc. etc.Very important to know what you are relying on, and anticipate how that might change over time. (e.g. storage increases)
If working in an environment with multiple funding sources, be very clear about what funding source is paying for what activity.If you are seeking funding (internally or externally) you must be clear about what that will AND WILL NOT pay for.Don’t find your ability to continue a core activity under threat because a funding source expires.What constitutes a core activity is for you to define.
Perhaps the most important cost question to consider is: what happens after the project when the money runs out?Incidentally, this is when you really find out who pays for what!
Ithaka summarise some of the ongoing requirements in a nice hierarchy.All projects or content will require a baseline of support.There may well be additional activities which will contribute towards the long-term sustainability of the content you produce.
A longer list of the kind of activities that may need to be resourced in the long term.Notice that actually producing content is a very small component.
To return to the “why” – sustainability is about impact.If you understand your users and deliver value for them then resourcing activities is an easier case to make.This is the flip-side of cost – what you achieve by making the investment.Impact measurement is, I would argue, still relatively immature. There are good resources out there and these are two of the best, but developing your own suite of metrics is important as you think about the longer-term nature of digitisation at your institution.
My contention, that the long-term is about an organisational attitude towards digital – how core are these activities to your mission and strategy?To what extent is your organisation willing to invest in skills and infrastructure to support digital?How much will your senior managers core-fund?
Building skills and keeping them.Able to focus on a known range of format types.You have time to implement processes and develop efficiencies.You have enough stuff to make this worthwhile.Maybe you are working in new areas and need to do R&D to work out how (probably unlikely).Perhaps it is just easy to make certain kinds of business case in your institution (e.g. staff budget underspends, or IT capital underspends)
Capital cost for equipment purchase, high-end equipment often comes with a yearly maintenance contract.Building skills, especially in volunteers, is relatively expensive, until you achieve efficient processes.Are you digitising a range of material that needs different equipment (flat bed, camera, specialist film, large-format, book cradles, etc.)Are you really digitising enough stuff on an ongoing basis to justify the investment in capacity?How will that change over time?
You need skills and experience, and you need them now.You have a range of different collections and can’t hope to implement skills and equipment for them all.You have a relatively small amount of time or material. Or, you have a lot (out-sourcing can work in both situations)You can specify the standards for handling and production very clearly, rather than expect a long process of investigation (most likely you will need to do this up front if necessary)Perhaps it is easier to make the case for a capital investment rather than staff and equipment maintenance.PRAGMATICS!!
What support do you need to manage the suppliers?Can you write a good spec, if you leave something unclear or non-specific it is not the suppliers fault.Do you have to work with finance dept? How do you locate and assess suppliers?Can you specific quality criteria and how will you measure them?Are you likely to need to source suppliers regularly, how does this work for procurement?How is your collection best divided, and how do you match suppliers?
LSE: no space, diversity of content, now making space but still planning to bring in external suppliers to fill it, as that is how we expect to achieve the greatest efficiency.
Quick pause, here’s a pretty picture to look at.And now I’ll attempt to summarise all of this in the last couple of minutes.
There are as many cost models as people who have ever tried to cost a digital lifecycle with any degree of seriousness.The learning – the models depend almost entirely on local assumptions.The lesson – know your institution, your funding sources, the responsibilities, and the business cases you will have to make.