Experiments in Crowd-funding Community Archaeology

Abstract This article reviews existing case studies in the ‘crowd-funding’ of community archaeology, as well as offering preliminary results from a small-scale experiment conducted alongside the wider crowd-sourcing efforts of the MicroPasts project (http://micropasts.org). In so-doing, it also considers the possible role of a hybrid reward- and donation-based model for micro-financing collaborative archaeological research. The article concludes with a summary of the key lessons drawn from experiences of crowd-funding archaeology so far, and highlights their particular implications for community archaeology.

and their relevance to community archaeology. In recent years, UK archaeological societies and other groups that are active in the preservation and research of their local heritage have relied greatly on grant funding to support their activities (Richardson 2013, 2;Thomas 2010). In particular, they have been largely subsidized by the Heritage Lottery Fund (HLF), to the extent that community archaeology has actually witnessed a substantial growth since the inception of the HLF in 1994 (Moshenska et al. 2011, 94;Simpson and Williams 2008). This development has not been curtailed by the recent economic crisis, which, instead, incentivized the purchase of lottery tickets and resulted in a £25 million increase in HLF income from 2009 to 2010 (Lottery.co.uk 2010).
At a time when financial hurdles compelled commercial archaeology companies, heritage organizations, and local authorities to scale back their education and outreach efforts (see the discussion in Thomas 2010), the HLF was essential in sustaining communities' manifold interactions with archaeological heritage. In addition, it enabled a large training programme to build community archaeology skills across the heritage sector through the 'Skills for the Future' scheme, managed by the Council for British Archaeology between 2011 and 2014 with further support from English Heritage, Cadw, and Historic Scotland (Sutcliffe 2014;Thomas 2014). As the HLF evolved into a primary financing source for community archaeology, however, competition to obtain HLF money exploded. Data from the last two financial years shows that the total funding amount applied for in 2013-14 grew 18 per cent from 2012-13three times HLF's income (Heritage Lottery Fund (HLF) 2014, 3). In this context, researchers have devoted particular attention to understanding the unique contribution that the internet and digital technologies can offer to enable mixed financing models (e.g. Rotheroe et al. 2014; Röthler and Wenzlaff 2011 1 ) useful to aid the sustainability of the heritage sector and of the arts and culture 'ecology' (e.g. Doeser 2014) more generally. A recent survey has revealed that, amongst all non-grant finance methods (including debt, equity, and alternative finance; e.g. Rotheroe et al. 2014, 6), crowd-funding has been receiving the greatest interest from heritage organizations, sometimes to the benefit of community archaeology work as well (Rotheroe et al. 2014).

Introducing crowd-funding
Crowd-sourcing is the collection of information, services or funds in small amounts and from relatively large groups of people via the internet (e.g. Dunn and Hedges 2012). When crowd-sourcing allows attracting venture capital online, it is also known as crowd-funding (e.g. Agrawal et al. 2011;Lehner 2013Lehner , 2014. As highlighted by several commentators already, this fundraising practice has not yet been extensively researched, partly because it is still in its infancy (e.g. see Lehner 2014;Mollick 2013). So far, studies on this topic have emerged primarily in the fields of business and economics and, more recently, new media (Bennet et al. 2015). They have focused on examining the financing of creative projects (e.g. Belleflamme et al. 2010;Sørensen 2012;Ward and Ramachandran 2010), social ventures (e.g. Lehner 2013Lehner , 2014, and initiatives in the world of information and communication (e.g. Jian and Shin 2015). This existing body of literature has identified four possible models of crowd-funding, depending on the kinds of reward that backers obtain (or not) in exchange for their contribution. In the patronage (or donation-based) model, supporting crowds do not receive anything back and act in ways akin to philanthropists, whereas in rewardbased crowd-funding they essentially pre-buy the product of the venture they are financing or a related outcome (Mollick 2013). In addition to these two models (the most common in social ventures), there is the lending model, where funds are given as loans 'with the expectation of some rate of return on capital invested', and the equity model of crowd-funding, which assigns equity stakes to investors (Mollick 2013, 3).
Research using data from some of the most popular generalist platforms and adopting a theoretical approach based on Bourdieu's forms of capital (Bourdieu 1986) has been key to revealing and explaining the two-tier mechanism for successful reward-based crowd-funding. Results show that funding goals tend to be reached when the project proponent is able to mobilize an initial group of donors amongst his/her own social network, and to motivate this first tier of supporters to leverage their own social capital, with the ultimate effect of involving a larger second-tier of people who are unknown to the campaigner (Lehner 2014). In this process, 'entrepreneurs' with larger and, crucially, more widely-spread networks are more likely to succeed than those with more tightly-knit ones (Hekman and Brusse 2013). Quantitative studies have identified other predictors of success, including whether 'entrepreneurs' have pledged to other projects in the past, thus inducing a sense of obligation towards them; the kind and frequency of updates provided on the progress of the campaign; the length and detailed nature of project descriptions; and the presence and quality of videos illustrating the bids (e.g. Colombo et al. 2014;Hekman and Brusse 2013;Xu et al. 2014;Zheng et al. 2014). A recently published report prepared for the UK-based innovation charity Nesta has also showed some of the key differences in the use of the four crowd-funding models mentioned above (Baeck et al. 2014). For example, it highlighted that donation-based crowd-funding is decreasing in frequency by comparison to the reward-based kind, but nonetheless allows the raising, on average, of slightly higher sums (c. £6,100 versus £3,800 by one estimate), and is linked to donors' volunteering for the project they backed (Baeck et al. 2014, 71, 85). While we could expect some of these insights to be valid for all kinds of crowd-funding ventures, there remains the need to verify how they relate to the financing of cultural heritage projects specifically. 2

Crowd-funding community archaeology
Even though they have not been formally reviewed until now, there are a number of cases where crowd-funding has been used to support community archaeology projects. The appeal must reside, at least in part, in the fact that it does not require the development of complex business plans or lengthy grant-funding proposals. These qualities make it especially palatable to organizations, individuals and community groups who pursue social and/or heritage ventures and, understandably, often do not have a background in finance (Lehner 2013, 290). So far, examples of crowd-funding community archaeology initiatives have fallen into two main categories. There is a smaller group of campaigns which have been run on generalist platforms such as https://www.kickstarter.com, https://www.indie gogo.com, or http://www.crowdfunder.co.uk. Projects on these platforms have focused primarily on conservation (e.g. site preservation) or education (e.g. supporting young people's learning about urban archaeological heritage, building lab facilities, or creating toolkits for outreach activities), whereas bids proposing to conduct archaeological research through partnerships with community groups are much less frequent. In contrast, a second group of projects has sought financial contributions via one of the two (to our knowledge) heritage-themed crowd-funding platforms that existed prior to the development of https://crowdfunded.micropasts.org/ (MicroPasts crowd-funding). These are http://commonsites.net/ (referred to here as CommonSites) and http://digventures .com/ (DigVentures). CommonSites and DigVentures (DV) share a broad mission of promoting, respectively, 'sustainable heritage practices' and 'sustainable archaeology and heritage projects'. A key difference between the two is that although both aim to be international and all-encompassing in the types of activities they support, so far DV has concentrated principally on financing archaeological excavations in the UK (projects elsewhere are planned for the future; Westcott Wilkins, personal communication). Between 2012 and 2014, DV has attracted 'seed' funding for three ventures of this kind, raising over £64,000 (Westcott Wilkins, personal communication). In contrast, as of January 2015, only six out of 62 projects hosted by CommonSites have a community element and only one of them promises to carry out collaborative research (the UK-based Archaeology in Telford Park project has been online since 2012, but has not to date reached its funding goal of £18,000).

The MicroPasts experiment of crowd-funding
This necessarily brief overview also highlights at least two characteristics that are as yet largely missing from existing crowd-funding efforts, but that we consider critical in the context of contemporary archaeological ethics, theory, and practice. The first is the need for stronger co-design and co-delivery of research between academics based in institutions (e.g. universities and museums) and interested community groups, whether the latter live in geographic proximity (communities of place; Scott and Johnson 2005) or simply share common interests online or offline. The second is the importance to shift the focus from 'digging' as archaeology, in favour of non-destructive 'light methodologies' (Vannini 2007;Vannini et al. 2014) such as buildings survey or online 'citizen archaeology' . The adoption of 'light' approaches allows for the creation of datasets that enable new statistical analyses and the re-interpretation of already known records at a larger scale, both geographically and chronologically (on the data deluge, see Bevan in press). The MicroPasts project is one attempt to encourage these two strands of public archaeology, and the discussion below summarizes its overall goals as well as a small experiment within it to crowd-fund community archaeology.

The MicroPasts project and crowd-funding model
MicroPasts is a partnership between the Institute of Archaeology, University College London (UCL), and the British Museum, to foster collaborations between institutions and members of the wider public (whether or not already organized in 'communities') to conduct research into the human past Bonacchi et al. 2014Bonacchi et al. , 2015. A main goal of MicroPasts is to generate open data that is useful for archaeological or historical investigations via the crowd-sourcing of individual tasks (http://crowdsourced.micropasts.org/). However, an additional experiment has been to crowd-fund research projects which involve (in both planning and implementation) mixed teams of traditional academics and volunteer communities. These projects have either been submitted to the site by existing partnerships or (potentially) will be designed by MicroPasts contributors online. Our focus here is on this experiment, which we will refer to as MicroPasts crowd-funding (https://crowdfunded.micropasts.org/).
The MicroPasts crowd-funding platform uses an open source framework that originated in Brazil as Catarse and developed further into the Neighbor.ly project. We chose PayPal as the payment gateway for a number of reasons; firstly its ubiquity and trustworthiness, secondly its ease of integration with the framework, and thirdly its reasonable cost, compared to other systems. Crowd-funding proposals can be submitted online, via the platform, and are vetted by the MicroPasts team to ensure that they meet the following requirements: (a) at least one institutional and one community coordinator is identified, so that the project outcomes will have value for both academia and communities; (b) the research project does not comprise excavation; and (c) the funding goal is set from a minimum of £500 to a maximum of £5,000. 3 Once a proposal is accepted, the proposing team is asked to finalize the bid and provide: a title for the project; a pitch line; a description of about 300 words (including research aims, context, and outcomes); an image; a short video (optional but encouraged); and an indication of the duration of the campaign. We recommend this to be 30 days to maximize chances of success, in line with previous research and anecdotal evidence (Mollick 2013). Ninety-nine per cent of the funds raised through MicroPasts crowd-funding are transferred to the project coordinators, regardless of whether the overall funding goal has been reached and, under our current protocols, communities can use any residual sums to support their day-to-day activities. MicroPasts retains the remaining 1 per cent for maintaining the platform. Proposals which have an excavation component are not accepted because, as noted above, our aim is to facilitate 'light' methodological approaches to archaeological research. We also decided to complement (rather than overlap) with existing crowd-funding provisions such as DV, which currently has community excavations at its core.
We initially implemented a hybrid donation/reward-based model of crowdfunding where donors are not promised prizes commensurate to the amount of money they have contributed, but (as any other member of the public) they would eventually be able to access some of the project findings via the web, and/or participate in talks or show-and-tell days (either in person or by watching video-recordings of these events online). In this way, it has been possible to test the effectiveness of a model that was expected to attract four non-mutually exclusive categories of donors (reviewed and integrated below): (a) those who support the project proponents because they are friends, family, or colleagues; (b) people who want to see a research undertaking happen out of interest in the results; or (c) out of an interest in creating an opportunity for themselves to volunteer on the project; 4 and (d) those who back a venture that they more broadly consider worthwhile for society or their local community. The third category (c) can include both 'local' volunteers operating offline, and online volunteers, if the proposed project entails the development of crowdsourcing applications, which would be deployed via http://crowdsourced.micro pasts.org/. Testing this model allowed for the assessment of the extent to which the intrinsic desire of engaging with the process or the results of archaeological and historical research, and/or the belief that this research has value for society or specific communities, can be drivers for the micro-financing of community archaeology projects, compared to the motivation of supporting one's own network, which tends to be popular in straightforwardly reward-based crowd-funding (Baeck et al. 2014).
The public launch of http://crowdsourced.micropasts.org/ was on 16 October 2014, with three 'seed' projects about community archaeology, to which a fourth on community (oral) history was added soon afterwards. All of these projects initially opted for 60-day-long campaigns and proposed to conduct research concerning the medieval or post-medieval periods (up to the present day) in England, UK (see Table 1 for more details).
The proponents had primary responsibility to publicize their ventures amongst their own personal, institutional, and community networks, although these efforts were sustained via the MicroPasts social media accounts and through further Micro-Pasts team outreach activities. Such activities included direct emails to the volunteers involved with MicroPasts crowd-sourcing (for early results and preliminary observations of these, see Bonacchi et al. 2015) as well as to local newspapers, schools, and archaeological and historical societies. This was in addition to the campaign delivered jointly by the press offices of the British Museum and UCL to promote the MicroPasts crowd-funding platform as a whole. To keep donors and prospective supporters informed of the progress of their fundraising, project coordinators also had the opportunity to post updates on the site.

Results
From 16 October 2014 to 11 February 2015, 5 the web page http://crowdfunded. micropasts.org/ was viewed 2,633 times, with peaks on 17 October 2014, the day after the launch, and on 28 November 2014, when the oral history campaign went live (Table 2). This is a considerably lower number of page views than registered for the MicroPasts crowd-sourcing 'home' (http://crowdsourced.micropasts. org/) in its first four months of public activity (13,952 views from 14 April 2014 to 11 July 2014). The data perhaps underscores the difficulty, even for larger and well-known international organizations in the higher education and heritage sectors, to establish new fund-raising spaces online. In addition, and to confirm further the challenges highlighted above, the website appears to have been visited EXPERIMENTS IN CROWD-FUNDING COMMUNITY ARCHAEOLOGY by substantially fewer people outside the UK (83 per cent Europe, 68 per cent UK, 43 per cent Londonall percentages refer to the total of page views).
The projects received the majority of their contributions in the initial 60 days of fundraising and just a few more followed in the subsequent two-month extension, at the end of which none of the 'seed' campaigns had reached their minimum funding goals. However, as illustrated in Table 1, the London's Lost Waterway venture raised 65 per cent of the desired amount (£3,000), which was sufficient for the team to start their activities (these officially launched in February 2015). The other three projects collected small overall sums ranging from £176 to £290 (7-27 per cent of the original target). Closer examination of donors, donor motivations and individual contributions revealed the reasons for the significantly more positive outcomes of London's Lost Waterway. This information was gathered directly from the crowd-funding website (absolute numbers of donors and donations) and thanks to a short survey that was coded in to appear after a donation had been made, with the possibility, on the part of the contributor, either to skip the step or submit a completed form. The form asked four questions enquiring about the main reasons for supporting a specific project, 6 the relationship between donor and project coordinators, 7 the country and city (or place) where the donor lives 8 and his or her email address, to enable in depth follow-up questioning if needed. This two-step evaluative approach proved successful and allowed the collection of valid (usable) responses from 50 out of 73 contributors (not including project staff and campaigners; Table 3), resulting in a 68 per cent response rate, a remarkably high figure if compared, for example, with that of end-of-project surveys that were emailed to users of crowd-sourcing platforms (cf. 8 per cent response rate in Causer et al. 2012). Individual donations varied from a minimum of £1 to a maximum of £1,000, but 75 per cent of them was equal to or less than £25, and 50 per cent of the total amount donated across the four projects was below £10 (Table 4). Although they were for the most part 'small chunks' of money, as the canonical definition of crowdfunding would require (see above), they were not raised from a 'crowd' but, rather, from a relatively low number of UK-based donors (only eight donors were living in other countries -USA, Australia, New Zealand, or Italy), which was higher exclusively in the case of the London's Lost Waterway project (Table 1).
Survey responses from these donors were correspondingly few, but they still provide some qualitative insight. From an analysis of 46 relevant answers (see Appendix -http://www.maneyonline.com/doi/suppl/10.1179/2051819615Z .00000000041), it is possible to suggest three non-mutually exclusive kinds of donor agendas: a research agenda, a place-making agenda, and a social agenda ( Table 5). The first category was the most recurrent, even though respondents were more frequently interested in hearing the results of the projects than in participating in the research process itself, whether offline or online. The three community archaeology 'seed' ventures offered the latter option, which included a crowdsourcing component, and were advertised amongst MicroPasts crowd-sourcing volunteers. An email about the crowd-funding projects was sent to the 827 subscribers on the crowd-sourcing side of the MicroPasts website, 9 but resulted in only 5.8 per cent of addressees clicking through to the http://crowdfunded.micropasts.org/ link. This suggests strong separation of those interested in crowd-sourced data collection and crowd-funding projects. Thus, the initial idea of a joined up approach where crowd-sourcing leads to collaborative design of new projects that are then crowdfunded has not proved easy to implement so far, and the use of the MicroPasts crowd-funding website remains largely independent from that of the crowd-sourcing site and the forum.  Returning to the survey, a crowd-funder motivation to connect with a 'personal past' was common and underlying a commitment to improve one's local area by discovering and preserving its history (place-making agenda), and, often, also the desire of undertaking research (e.g. out of an interest in knowing more about one's own place of residence or in genealogy and family history). However, people donating online to enable research which was relevant to them personally, and in which they could potentially take part, remained inadequate to ensure the success of the campaigns. This was probably the consequence of two main factors. The first is that people supporting initiatives in their local area are not necessarily inclined to respond to an online crowd-funding model: for example, several of the 'seed' projects include many community archaeology members from older age groups (see also Thomas 2010, 23) who have generally lower engagement with digital technology, and, perhaps also lower levels of trust in online forms of payments (as suggested to us by some of the same community members). A second factor is that these campaigns all collected some donations offline (e.g. at community events) in tandem with online crowd-funding efforts and publicity, but only the London's Lost Waterway team submitted these donations via the MicroPasts crowdfunding platform. Finally, just a small group of backers donated because they acknowledged the wider value of the proposed research conducted: most saw limited relevance beyond the local scale, even though the project descriptions highlighted how they would be contributing to advancing knowledge of a certain period of history or a specific research theme. On the basis of this admittedly very small experiment, it seems difficult to leverage a pure donation-based model for crowd-funding non-excavation community projects. However, when tailored rewards were subsequently offered for contributors to The Origins and Administration of Anglos-Saxon Wessex, for example, they did not bring any significant change in the amount of donations that were raised. The London's Lost Waterway project suggests that a hybrid donation-reward model may work when: (a) it taps into a numerically larger (and probably, on average, demographically younger) community group, and (b) it relates to more recent times in history and/or a place-making component (in this case, documenting the foreshore before it is destroyed) that make the project more directly relevant to local as well as a few more international contributors. In addition, this 'seed' campaign attained more positive results because it interpreted crowd-funding as a means of focusing attention on the fund-raising venture and as a mixed method which could also solicit larger donations (a £1,000 sponsorship, for example) and offline contributions by community members during special events.

Conclusion
Our brief experience of seeking financial support through the MicroPasts crowdfunding platform has confirmed the importance of understanding how general funding methods must be tailored to specific areas of research or practice. The very low amounts of funding secured by three of the four 'seed' projects indicate that the crowd-funding model chosen by MicroPasts has not been an effective solution to aid small community archaeology ventures. However, the study presented here has identified and discussed some of the reasons for these failures, as well as highlighting possible ways of overcoming them. Since a difficulty emerged not only in funding projects not comprising excavation, preservation, or public interpretation elements, it is suggested that these components are, to some degree, moulded in the crowd-funding campaigns, which can nonetheless also retain 'light' methodological approaches. Furthermore, the hybrid activities of an initiative such as London's Lost Waterway seem a more viable mixed model of funding to pursue. Hence, for community archaeology and history projects, it might be more appropriate to consider crowd-funding as a catalyst for collecting various kinds of contributions. These may comprise those offered online by people within the network of the campaigners or by unknown individuals internationally, but also solicited sponsorships and offline donations (this has also been the choice of DV so far). While the latter do not technically need a crowd-funding platform, they are largely a product of the overall crowd-funding efforts.
A tailored interpretation of the 'traditional' donation-based model of crowdfunding along the lines described above seems a useful method to finance small (and 'light') community archaeology ventures, perhaps in tandem with traditional grant-financing. In this form, web-based micro-financing is more aligned with the rationale of social ventures, and its success tends to be associated with the support of a communal cause or the construction of a personal past than the commodification of research and heritage via their transformation into pre-purchased goods, presented as rewards. Such observation provides a counter-argument to some of the criticism directed towards crowd-based methodologies, described as promoting falsely democratic models (e.g. Gilge 2014;Harrison 2010), that impose institutional agendas and build knowledge by discarding the alternative views of minorities.

Acknowledgements
The MicroPasts project was funded for an initial period of 18 months (1 October 2013-31 March 2015) by the UK Arts and Humanities Research Council, through the themed call on Digital Transformations in Community Research Co-Production in the Arts and Humanities. We thank the project coordinators and community groups who have chosen the MicroPasts crowd-funding platform for their fund-raising campaigns, and all the donors who have supported these projects. Our thanks are also to Irio Musskopf, for helping to adapt the Neighbor.ly crowd-funding framework to suit the needs of MicroPasts, and to Jordan Hassell and Oliver Hutchinson for promoting some of the 'seed' projects. Finally, we are indebted to Lisa Westcott Wilkins and Meghan Ferriter, for useful discussion on the topic of crowd-funding cultural heritage, to the two anonymous reviewers, and the editors of this journal.
Notes 1 See also, for example, http://ec.europa.eu/culture/ news/2014/20140428-call-cro wdfunding_en.htm. 2 Publications about crowd-funding in the heritage and archaeology domain are virtually non-existent, to our knowledge. The topic is, however, addressed in the unpublished MA dissertation Koivisto (2014). 3 We introduced a maximum funding goal of £5,000 to contain risks of failures, since a recent report found the average amount raised by donation-based crowd-funding to be £6,102 and £3,766 for rewardbased crowd-funding (Baeck et al. 2014, 11). 4 Recent research has in fact evidenced how almost a third of donors in donation-based crowd-funding have also offered to volunteer on the project they supported (Baeck et al. 2014). 5 This is the period from the public launch of the USA; (c) Other [specify]; Please indicate the CITY/ PLACE where you live. Open answer. 9 These subscribers included registered members as well as the anonymous ones who had provided their email addresses via a survey form that popped-up after they had completed one task on crowdsourced.micropasts.org Bonacchi et al. 2015).