A Right Time to Give: Beyond Saving Time in Automated
Conditional Donations
Nicola J. Bidwell, International University of Management, Namibia and Northumbria University, United Kingdom,
nic.bidwell@gmail.com
Chris Elsden, Institute of Design Informatics, University of Edinburgh, United Kingdom, celsden@ed.ac.uk
Ludwig Trotter, School of Computing & Communications Lancaster University, United Kingdom, l.k.trotter@lancaster.ac.uk
Josh Hallwright, Oxfam Australia, Australia, joshh@oxfam.org.au
Sadie Moore, Oxfam Australia, Australia, sadie_moore@hotmail.com
Kate Jeite-Delbridge, Oxfam Australia, Australia, katedelbridge@gmail.com
Mike Harding, School of Computing and Communications Lancaster University, United Kingdom, m.harding@lancaster.ac.uk
Peter Shaw, School of Computing and Communications Lancaster University, United Kingdom, p.shaw@lancaster.ac.uk
Nigel Davies, School of Computing and Communications Lancaster University, United Kingdom, nigel@comp.lancs.ac.uk
Chris Speed, Institute of Design Informatics, University of Edinburgh, United Kingdom, c.speed@ed.ac.uk
John Vines, Institute of Design Informatics, University of Edinburgh, United Kingdom, john.vines@ed.ac.uk
Abstract
Smart Donations is a blockchain-based platform that enables users to program and set conditions on charitable donations. Users set up contracts
to donate in response to real-world events, e.g., whenever an earthquake is detected or an activist tweets about refugees. We designed Smart
donations with Oxfam Australia, trialled it over 8-weeks with 86 people and qualitatively analysed questionnaires and interviews about their
experiences. Temporal qualities emerge when automation enforces conditions that determine when donations are made. These contributed to
participants’ sense of immediacy in donating to humanitarian crises, ongoing involvement while awaiting conditions to be met, and awareness of
events that are usually unconscious. We suggest that automation can reveal diverse temporal registers, in real-world phenomena, sociality, morality
and everyday life, which contribute to experiencing a ‘right time’ to donate. Thus, we recommend adopting a sensitivity to right time in designing
for multiple temporalities in FinTech more generally.
CCS CONCEPTS
• Human-centered computing~Interaction design~Empirical studies in interaction design
Keywords
Temporal Design, Blockchain, FinTech, Philanthropic Informatics, Right time, Temporal marginalisation, Financialisation of natural
disasters, Automation, Programmable donations
1 Introduction
The focus on efficiency, convenience and saving time in visions of new data-driven technologies is often critiqued in HCI
research on Philanthropic Informatics and financial technologies (FinTech) [27, 32, 39, 44]. Studies repeatedly observe the
diverse and situated nuances of philanthropic work [33] and the artful and creative ways that people manage and socially
interact around digital and analogue money [72, 94]. Complex relations also emerge in automated, data-driven personal
transactions; for instance, communication and self-expression appear as important as financial efficacy when people programme
rules and intentions using a mobile banking app [18]. In this paper we explore people’s experiences in programming conditions
that determine their donations to charity. We suggest temporal qualities, that emerge when automation enforces these
conditions, shape experiences of giving that are not limited to saving time.
We analyse user experiences in a trial of a new platform that enabled participants to program and set conditions for charitable
donations. Unlike giving ad-hoc in response to a fundraiser or regularly via a monthly bank transfer, donations are automatically
driven by data about specific phenomena. Smart Donations, a blockchain-based platform, offers users ‘contracts’ that donate
money to certain causes in response to real-world events. For example, participants could donate when an earthquake is
detected or whenever an activist tweets about refugees. If the conditions the user sets are not met by the time the contract
expires funds are returned to them. Our 8-week trial of Smart Donations produced several distinct contributions, including the
technical implementation [90], platform governance, interface design and the choice of data to drive donations. This paper
focuses on the temporal qualities of users’ experiences that emerged when automation, based on data-driven transactions,
enforces their financial intentions. Programming a smart donation does not simply mediate if a donation is made, but when it is
2
made. Temporality is a well-recognised feature of people’s experiences of technology [73], and designers are increasingly urged
to recognise “pluritemporality” [66] and that time is relational and differential (e.g. [48, 75]. However, existing data-driven
technologies neglect diverse experiences of time and little prior work has explored temporality in the context of FinTech or
Philanthropic Informatics.
Rather than prioritizing efficiency, Smart Donations enforces exactly when transactions occur. Participants in our trial set up
‘smart contracts’, on a mobile app, to donate to various programmes run by Oxfam Australia. The contracts used data about
relevant real-world phenomena to determine if and when to release pledged funds to a specified programme. In addition to
temporal features integral to blockchain technology, contracts were triggered at different thresholds and frequencies and over
different durations, based on the conditions that participants set. We surveyed 86 participants before and after the trial, collected
analytics of their app usage, interviewed 15 of them afterward, and then analysed their experiences of the temporalities that
emerged in the trial. Some felt an immediacy in the tie between their donations to humanitarian crisis and data about
phenomena that cause them; some felt an ongoing involvement in causes after setting criteria; and many became more aware of
the social, political, geological and meteorological phenomena that triggered donations.
Participants interviewed in our prior study about the concept of ‘programmable donations’ speculated about the challenges of
negotiating temporality and predicting and waiting for contract conditions to be met [21]. They suggested that while upfront
commitments and rational links between donations and events might feel worthy, the ‘warm glow’ felt in donating would be
delayed or absent, and the uncertainty about a conditional donation occuring may induce anxiety. Our analysis of deploying a
real-world trial of the fully functional Smart Donations app, however, uncovered other relations between felt-experiences, moral-
reasoning and the temporalities of users who pledge donations, of phenomena that create circumstances that need support and
of a charity’s response. Thus, we suggest the relations that emerge in using Smart Donations contribute to different experiences
of a right time to give.
The term, right time, was first introduced to HCI by Aboriginal authors to describe the appropriate timing of cultural practices
when ecological, ancestral, metaphysical, meteorological and other phenomena converge [10], however non-Aboriginal people
also relate to it [89, 99]. We use the term to speak to the multiple ways people experience and interpret time and reconcile
different temporal orders. Right time involves sensing and responding appropriately to the timing of events and may help us
design to support an awareness of diverse temporalities in environmental and social phenomena, emotions and moral
reasoning. Thus, we extend literature about time in HCI and the way algorithms and data-driven and automated technologies
reify particular temporal arrangements [75]. We contribute empirical insights about the effects of synchronising donating with
real-world events on felt-experience, including senses of immediacy in contributing to crisis response and ongoing involvement
in situations far-away, and awareness of events that are usually unconscious. Our analysis highlights, firstly, the role of
automation in juxtaposing diverse temporal patterns and, secondly, how data-driven automation in philanthropy contributes to
experiences and reasoning about right times to give, beyond productivity or efficiency.
2 Related work
Participants’ experiences of automation, in the trial, reveal different “temporal registers” or general rhythms [38] in real-world
phenomena, in sociality, morality and everyday life. Their experiences inspired us to reflect on how reconciling different registers
yields a right time to give. Thus, we begin by anchoring our perspective in the diverse literature on temporality in HCI, and then
connect this to automation and studies in Philanthropic Informatics and money, on which our research builds.
2.1 Right Time in HCI
In media studies, Bucher proposes that by sorting, filtering, ranking, curating content and aggregating data on people’s
interactions, algorithms personalise a “right-time” for a user to experience media [12]. She compares this right-time to ‘kairos’, a
Greek concept that refers to an opportune time to say or do something or an unpredictable moment that presents an occasion
that must be seized. Kairos has inspired work in HCI; the Kairoscope, for example, is an automatic scheduling system that aims
to reduce users’ stress about when exactly events will happen by both optimizing their time use and assigning precise times only
as appointments approach [51]. Kairos contrasts with the ‘right time’ described by Aboriginal authors in HCI, who undertake their
cultural practices when certain tangible and intangible circumstances converge [10]. Non-Aboriginal people also associate right
time with states when an external temporal regime is not imposed, such as sensing the moment to lower the heat in cooking a
familiar recipe or “me-time, where I can be as by myself” [99].
Like Taylor et al, we interpret right time as “a set of converging circumstances that constitute “the time” for happenings to take
place” and recognise that practices unfold in ways that are particular to a situation [89]. Importantly, we account for the way
Aboriginal authors relate right time to being aware of ecological and moral, as well as personal and social, cues by observing the
circumstances involved [10]. This differs from Bucher’s understanding of right-time as a “personalized moment instigated by
aggregated individuals, and fuelled by the business models of platforms” [12]. We propose feelings and reasonings that relate to
right times to give go beyond optimizing time use and are experienced through the intersections and juxtapositions of diverse
temporal patterns and orders. Thus, next, we describe how digital devices are often framed by less nuanced beliefs about saving
time, by doing things faster and more efficiently. Then we refer to some, of many, studies to suggest that optimising time in
certain ways discounts the value of adapting to different temporalities.
3
2.1.1 Saving Time in a Pluritemporal World.
Our experiences of time are simultaneously quantitative, objective, subjective, situated and socially constructed, and we assign
many meanings to these experiences [66]. Negotiating such “pluritemporality” [66] and multiple temporal orders and registers is
vital to collaboration. Consider, for instance, digital humanitarians distributed across many countries applying their various
expertise to respond to disasters [58]. They devote effort not only to managing many temporal orders in incoming streams of
information, metadata structures and systems that encode and display time in various ways but also to reconciling diverse
personal and cultural representations and experiences of time [58]. Despite this diversity, technological innovation and digital
devices are often framed in a “discourse of speed” that emphasises progress and busyness [95]. Technologies and modern
cultures of instantaneity [48] and hyper-productivity [95] have co-evolved, and are critiqued for affecting our well-being, and
senses of rush and lack of control [46, 81]. Such critique has prompted approaches to Temporal Designs [75] that encourage
more flexible and inclusive engagements with time, a more “human experience of time” and reflection or experiences over longer
durations linked to organic rhythms or different metaphors (e.g. [24, 31, 51, 60, 61, 62]). We often assume determinism in the
way technologies disconnect people from ‘natural’ lifestyles [99], however, we have also honed skills to coordinate our activities
with different measures and tempos [36, 63] and often only notice incongruence when one set of temporal registers intrude on
other obligations and desires in our lives [9].
Certainly, our experience of “digital time”, as we interact with technology, differs from chronological time and the social and
biological rhythms of everyday life (e.g. [99]). For instance, the flow of continually changing, but constantly accessible, digital
information contributes to the “condition of immediacy” [48], in which we interact with asynchronous information as if we are in
the same temporal register as it. This includes, for instance, the experience of “nowness”, when we treat social media posts as if
they report on current events and imagine them in an unfolding story with a past and future [34]. Technology use, social
conventions and temporal design entwine to uphold experiences of immediacy and nowness. In fact, while they seem distracting,
Lindley [48] suggests that many technologies do not set a pace of interaction but flexibly fit into our rhythms and afford an
experience that is both interruptible and expandable into unanticipated gaps in schedules (e.g. [37]). Yet, priorities for
productivity and efficiency often foster a need to fill this “plastic time” meaningfully [99].
2.1.2 Beyond Efficiency: Awareness in Reconciling Temporalities.
It's common to perceive the pace of modern life as accelerating, however studies show that we do not have less leisure time or
more things to do than people 80 years ago (see: [48]). Rather, we make more effort to coordinate our daily lives because work
patterns are more heterogenous and there are fewer fixed institutional events, such as standard mealtimes or ‘paydays’. We can
feel overloaded and busier by trying to stay “on the same page” in “collective time” [99] with our family and friends [48, 99], such
as when we enact our togetherness across the temporal fractures in our lives in an endless trace of WhatsApp messages [64].
Our temporalities are interwoven and interdependent, and some cultural practices more than others explicitly express the
inherent inseparability of time and human relationships. Work in ICTD, for instance, observes that design abstractions that
prioritise schedules and punctuality disrupt local temporal orders and registers in some societies (e.g. [9]). Yet participation in the
same temporal orders is distributed unevenly according to a society's power relations [80]. Some rhythms are prioritised in
divisions of labour, for instance those working on early morning or night shifts support those who work ‘9-to-5’. Tensions emerge
when algorithms designed to match supply and demand, such as timebanks and on-demand service platforms, do not account
for the way time is differential and the extra work some people must do to coordinate with others (e.g. [13, 83]). In fact,
automated on-demand services contribute to, and normalise temporal orders that marginalise ‘just-in-time’ workers. Consider
how Uber has only slightly increased the overall volume of taxi journeys [30] but significantly increased complaints to non-Uber
taxis [97].
In response to their marginalisation, platform workers opportunistically reconstruct their temporalities when trying to meet the
demands imposed or behave subversively [13, 53]. At peak hours, for instance, auto-rickshaw drivers in India ignore pick-up
requests from an Uber-like app because riders renege on bookings when drivers do not reach them fast enough, often without
cancelling [1]. Automation can undermine the transient awareness that enables people to coordinate. Compare automated ride-
sharing platforms, for instance, with the uninterrupted tempo achieved in Namibia’s self-organising, shared taxi system, where
drivers’ and riders’ mutual awareness enables drivers to keep all the seats in their taxis full and riders to hail an affordable ride to
their destinations from the roadside [41].
Indeed, successful collaboration depends on looking beyond individual people’s efficiency and reconciling the many rhythms of
different roles, jobs, identities, life-stages, organizations and infrastructures [6, 38, 45, 55, 78]. Jackson et al.’s study of scientific
collaboration highlights the work people perform to align to temporal registers in nature and society, such as circadian, seasonal
or perennial rhythms and episodic events [38]. The sense of right time amidst all this temporal diversity is less about optimising
time and more about noticing, and adapting to, the many orders and registers involved.
2.2 Time, Money and Philanthropy
All our interactions with money have temporal qualities whether spending to a budget, saving for the future or splurging bonus
income, and FinTech and banking applications often track and predict personal spending, over time, to support our co-ordination
of money [39, 47]. However, “moneywork” [72], or the appropriate ordering of different aspects of financial interactions that make
transactions technically and socially effective, is oriented by more than efficiency, speed and economic rationales [39, 94]. While
being able to pay at a particular moment enables successful transactions, we manage our money in “artful” and “creative” ways
[94] and the right time to pay is deeply situated. For instance, members of low-income households carefully delay or prioritise
paying bills to maintain their livelihoods and reputations [65, 91]. Enforcing a commitment to pay through technologies, such as
4
blockchain-based contracts (e.g. [56, 86]), must then account for diverse factors. Thus, next, we introduce literature on some of
the consequences of automation and conditionality on people’s financial practices before outlining some temporal issues in
philanthropic informatics.
2.2.1 Blockchain and Automated, Conditional Financial Interactions.
Rules and conditions have long been associated with different forms of “special monies” [44, 100], such as designating cash for
specific purposes in household budgets or restricting how food stamps can be spent. Recent attention has turned to
programming money and financial systems for certain uses and automatically enforcing payment [18, 21]. We are becoming
accustomed to an increasing range of interactive and digital representations of money such as virtual currencies, mobile money,
cryptocurrencies [17, 23, 40, 43, 96]; to financial services that depend upon transactional data [59]; and smart, autonomous and
“self-driving” money that adheres to sets of automated rules and procedures [52] and is already used in financial trading [29, 50].
Complex relations emerge between people and money in automated, conditional, data-driven transactions. Consider how
agency and expression are as important as efficiency when users of a mobile banking app programme rules and intentions into
their money, such as putting some into a savings account whenever it rains [18]. Consider also how conflicts emerge between
users’ immediate and more analytical thoughts about payment and the consequences of their choices when using a financially
autonomous coffee machine [87]. Building systems such as coffee machines and Smart Donations on a blockchain makes
automated rules and conditions immutable. Written in ‘smart contracts’, the code governing these conditions is hard-wired; the
consensus process and distribution of blockchain technologies ensure that protocols and transaction histories are transparent
and not easily changed [19, 77]. This immutability is important to trusting a financial escrow’s operation but also entails a pre-
determined approach. Prior research on interacting with blockchain technologies has focused on trust (e.g. [25, 42, 49, 79]),
conditionality (e.g. [21, 76, 86]), contracting (e.g. [56, 22]) and governance (e.g. [14, 57]); however, little has examined the
unique temporal experiences made possible when blockchains and escrows preserve and enforce past intentions.
Blockchains have other potential implications for temporal realities that motivate exploring user experiences with them in relation
to real-world events. ‘Blocktime’, or the time that it takes a certain number of blocks to confirm, is a separate time schema from
the time of events and temporal registers that shape fiat currencies. Manipulating the differential between these schema, by
configuring the conversion metric or the blockchain event trajectory, can ensure they reinforce each other or create more time by
arbitrage [84]. Thus, contracts could include blocktime offsets, for example, to increase the competitive advantage of algorithms
that financialize real-world events (e.g. [29]). This prompts considering the values in relation to blockchain and time.
2.2.2 Philanthropic Informatics, Data-driven Giving and Time.
HCI research on “charitable technologies” [28] encompasses organisational management [11, 33], delivering aid programmes [3,
35, 58, 92] and supporting fundraising and donating (e.g. [26, 82, 88, 98]). Researchers in the emerging field of Philanthropic
Informatics [92] have also argued that increased datafication and metricization can disempower non-profit organizations [11],
neglect the situated care entailed by their work [33] and, meanwhile, mediate donors’, stakeholders’ and beneficiaries’ values
[20, 93]. Extending this field to consider data-driven technologies in fundraising from individual donors, a co-speculative interview
study examined the potential of a ‘programmable donations’ service [21]. The study proposed that identifying data and conditions
suited to trigger a donation will present challenges but did not explore the temporal qualities that might emerge.
Indeed, given the improvements to efficiency and speed envisioned by Philanthropic IT [32], and often sought for responses to
humanitarian disasters, studies on the temporality of technology-mediated giving are strikingly absent. Nonetheless other
research shows temporality affects the complex feelings and morality associated with giving. Empirical studies suggest eight key
mechanisms drive us to give to charity, ranging from awareness of need and altruism, through to reputation and psychological
benefits [7], and that ambiguities arise when we are asked to give, for instance, faced by a fundraiser we might dislike being
asked or just dislike saying no [4]. Ambiguities and moral contradictions are also affected by elapsed time, between deciding in
advance to give and then actually giving. Delaying the transaction of funds, after committing to give, can avoid the guilt of saying
no, but also reduce the immediate feeling of a ‘warm glow’. While some people who pledge may renege later, others will give
more and the timing of pre-emptive thanks, after pledging, also affects if and how much people give [4]. In contrast with studies
of temporal effects on pledging, Smart Donations prevents reneging because it holds funds ‘in escrow’ after they are pledged
and releases them only when the contract ends. Thus, enforcing commitments and automating conditions introduces novel
temporal situations for charitable giving and opportunities to explore the experience of donations configured to occur at a specific
time.
3 Smart donations and Time
Smart Donations was developed as part our team’s research about blockchain technologies in the so called ‘first mile’ of
charitable giving, rather than for coordination, transparency or trust in charities (e.g. [15]). It uses an automated escrow to
determine if and when users make donations, not to track, record or disburse donations [3, 16, 74]. Some of the authors
conceptualised the platform in 2018, while establishing a partnership with Oxfam Australia and undertook fieldwork and
stakeholder engagement to explore its potential for fundraising. Employees of the charity felt that conditional, automated giving
could attract new donors to certain causes and sought to explore new user experiences of the many different conditions that
donors could set-up, not just match donor’s money to a specific need. While not the only way to deploy conditional giving,
blockchains offer an enhanced level of trust because they support distributed actors in transparently enforcing a set of rules and
governance of each transaction [90]. Thus, subsequently we worked together to create a broad range of content to implement in
a trial.
5
3.1 The Smart Donations Platform
The reference architecture for Smart Donations comprises three layers: (i) A smart contract compatible blockchain, (ii) off-chain
components, and (iii) and a mobile app (see: [90] for technical details). The distributed ledger, the blockchain, provides the run-
time environment for smart contracts, containing the logic for the value exchange and escrow and. For the trial we
operationalised the Smart Donations on a private, permissionless Ethereum network. External ‘data providers’, or oracles,
connect to the system via a secured Representational State Transfer API (RestAPI). The off-chain systems, for privacy-sensitive
user data of donors and beneficiaries, are hosted on traditional databases with the charities involved and we implemented their
functionalities as custom Python backend and SQL databases. The mobile phone application supports the two most popular
smartphone operating systems using cross-platform technology (Apache Cordova and Ionic 4).
The mobile app presents templates for ‘contracts’ that allow users to set up conditions to trigger a donation from funds that are
held ‘in escrow’ and released to the charity only when the conditions are met. Every template is a dedicated ‘smart contract’
deployed on the blockchain that records the donor's pledges. The blockchain assures the terms of the escrow cannot be
changed. By choosing a pre-set ‘smart plan’ or creating a custom contract, users can pledge funds providing there are sufficient
in their personal wallet. After ‘signing’ the contract and a 5-min ‘cool-down’ period, when a commitment can be withdrawn, the
pledge is being ‘sealed’. Once sealed, a donation cannot be altered and will be automatically issued if the set conditions were
met. Each contract expires after a defined duration, configured as 1-14 days in the trial (Tables 1 to 5), and returns any
remaining funds to the user’s wallet. App notifications inform users in real-time when their contract 'seals', fulfills or expires, and
when conditions are met.
An ‘active’ contract relies upon data from an authorised external data provider to ensure its conditions are met, for instance we
used the US Geological Society (USGS) for Earthquake Insurance contracts (Table 4). The contract releases a predetermined
amount of funds when a condition is met, such as when an earthquake above the threshold magnitude occurs in a selected
place (e.g. Figure 1). ‘One-off’ donations release all funds in the escrow if the condition is met, whereas ‘recurring’ donations
release funds whenever the condition is met until either none remain in the escrow or the contract expires. While the return of
funds underpins Smart Donation’s conditionality, configuring how much and when donations are made has more effect on a
contract’s temporality. The exchange of fiat money (i.e. AUD) a digital asset (i.e. a crypto- currency or a crypto token) is crucial in
order to manage, transfer and seal funds in self-governing digital trust funds underwritten by smart contracts on the blockchain.
To avoid interfacing with payment systems, we provided digital vouchers to participants in our trial, which we managed in a
dedicated token contract following the EIP-20 / ERC-20 standard. We settled the balance after the trial with research funds.
Figure 1: Screenshots of one of the contracts in the Smart Donations app. Each project included key information and
links about the donation and options to configure particular pre-set or custom conditions.
