Screening the Future 2012 Pause Play and Press

Screening the Future 2012: Pause, Play, and Press Forward Los Angeles, May 21 -23, 2012 Total Cost of Preservation Cost Modeling for Sustainable Services Stephen Abrams Patricia Cruse John Kunze University of California Curation Center California Digital Library

Goals n Understand costs in order to plan for and implement sustainable preservation services n Investigate the possibility of paid-up pricing in order to address Boom-or-bust budget cycles ► Fixed-term, grant funded projects ► rt po al re Fin Source: www. sharedidiz. com/

Meeting the expectation n Corner store ► n ≤ $ 100/TB Commercial cloud ► ► ► n Fry’s Amazon Apple Dropbox Google Microsoft $1, 000/TB/year $2, 000/TB/year $1, 980/TB/year $1, 200/TB/year $ 500/TB/year Preservation repository Princeton Data. Space $6, 000/TB for “forever” ► USC Digital Repository $1, 000/TB for 20 years ►

Prior work n Nationaal Archief (2005) } } Identification of granular cost components http: //www. nationaalarchief. nl/sites/default/files/docs/kennisbank/codpv 1. pdf n LIFE (2008) http: //www. life. ac. uk/ n KRDS (2010) http: //www. beagrie. com/krds. php n Data. Space (2010) http: //arks. princeton. edu/ark: /88435/dsp 01 w 6634361 k n Assumption of annual decrease in aggregate cost, i. e. , discounted cash flow (DCF) Jean-Daniel Zeller (2010) “Cost of digital archiving: Is there a universal model? ” 8 th European Conference on Digital Archiving, Geneva, April 28 -30, 2010 http: //regarddejanus. files. wordpress. com/2010/05/costsdigitalarchiving-_jdz_eca 2010. pdf n Rosenthal (2011) Critique of DCF approach http: //blog. dshr. org/2011/09/modeling-economics-of-long-term-storage. html

Cost model components �System, composed of various �Services providing curation function, running on �Servers, deployed by �Staff, in support of content �Producers, who use �Workflows to submit instances of �Content Types, which occupy �Storage, and are subject to ongoing �Monitoring and periodic �Interventions; all subject to managerial �Oversight

Total cost of preservation Total cost to service provider Fixed cost of oversight Number and unit cost of Interventions Fixed cost of System component subsumes Services System and Servers Number and unit cost of Producers Number and unit cost of Workflows Staff costs are subsumed by other components Unit cost and number of Content Types Number and unit cost of Monitoring Number and unit cost of Storage

Cost to a single producer n Cost of the Archive, Workflows, Content Types, Monitoring, and Interventions are “common goods” Equally beneficial to all Providers ► Properly apportioned across all Providers ► Total cost attributable to a given Producer Number of Producers Unit cost of a Producer Number of Storage units attributable to Producer

Price models n Pay-as-you go pricing Annual billing cycle ► Price = cost of providing curation service to a given Producer ► Only viable if Producers have predictable and reliable sources of funding ► ● n Any interruption in funding, and therefore service, may result in irretrievable data loss Paid-up pricing ► Assumption of an annual decrease, d, in aggregate service cost, and investment return, r ● In economic terms, this is a Discounted Cash Flow (DCF) or Net Present Value (NPV) model

Pricing over time $16, 000 (1–d)t discount factor Cost ($) $14, 000 Cumulative pay-as-you-go G (T ) Discounted pay-as-you-go $12, 000 G (T, d ) $10, 000 G ( , d ) Discounted pay-as-you-go $ 8, 000 $ 6, 000 $ 4, 000 $ 2, 000 $ 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Year (T)

Pricing over time $16, 000 (1–d)t discount factor Cost ($) $14, 000 Cumulative pay-as-you-go G (T ) Discounted pay-as-you-go $12, 000 G (T, d ) $10, 000 G ( , d ) Discounted pay-as-you-go $ 8, 000 $ 6, 000 $ 4, 000 Paid-up price, for T (1+r)t investment return $ 2, 000 $ 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Year (T) F (T, d , r) Paid-up price, for F ( , d , r)

Predicting the future is hard n TCP modeling is dependent on the predictive reliability of d and r www. flickr. com/photos/mcgraths/3248483447

Predicting the future is hard n Model the risk ► n Recalibrate the model ► n Add a risk premium to the modeling equations Periodically reset the values of d and r (and G and. F ), based on contemporary conditions, to be applied prospectively Bound the uncertainty ► Stochastic modeling to determine the probability distribution of possible outcomes

Hybrid price model n Distinguish between costs that are (relatively) easy to quantify and forecast, and those that aren’t ► Use the paid-up model for the former and pay-as-you-go for the latter Easy Difficult Archive Intervention Producer Workflow Content Type Monitoring Storage

Hybrid price model n Distinguish between costs that are (relatively) easy to quantify and forecast, and those that aren’t ► Use the paid-up model for the former and pay-as-you-go for the latter Easy Difficult Archive Content Type Producer Workflow Storage Monitoring Intervention

Preservation forever n Some things are intended to last forever… Source: John Church Company Source: United Artists

Preservation forever n Some things are intended to last forever… ?

Preservation for … n A fixed term – 10 years? 20 years? – may be appropriate for much content ► Give content an opportunity to prove its worth, as evidenced by someone’s commitment to pay for its subsequent preservation

Transparency and opportunity n Possible outcomes… ► We overestimate our costs and collect too much ● ● ► Fund a higher level of service Refund some portion We underestimate ● ● ● Ask for additional funds Lower service levels De-accession content – but at least it was preserved up to that point and had a chance to prove its value, and gain an advocate

Conclusions n Different customers have different funding capabilities ► n Any price model is based on an idealization of the real world ► n Assumptions matter Understanding all of your costs is a precondition to a policy decision to recover all or part of those costs ► n Flexibility in price models is important Cost accounting is difficult If investment return and discount factor can be reliably projected, DCF can be used to model of long-term costs ► What if not?

Conclusions n Even if we don’t have a perfect model, we need to move forward now with a “good enough” model Source: Getty Images

For more information Total Cost of Preservation: Cost Modeling for Sustainable Services http: //wiki. ucop. edu/display/Curation/Cost+Modeling UC Curation Center http: //www. cdlib. org/uc 3@ucop. edu Stephen Abrams Patricia Cruse Scott Fisher Erik Hetzner Greg Janée John Kunze Margaret Low David Loy Mark Reyes Abhishek Salve Joan Starr Tracy Seneca Carly Strasser Marisa Strong Adrian Turner Perry Willett Source: Getty Images