Service definition

To implement such a service, we registered Empath UK as a company limited by guarantee, and as a registered charity with the Charity Commission.¹Companies House registration no. 7041985; Charity Commission registration no. 1135025.  We describe in this section how we designed a working version of each of the components given above, the initial difficulties we encountered and how we addressed them.

Donor presentation

We considered several options for categorising charities, and looked at existing work in the same area.  In particular, classifications exist in charity law and in some of the services already described. One of the uses for the charity law classification is to make a clear determination of what activities can be considered charitable, and consequently whether an organisation can register as a legally-recognised charity.  On the other hand, classifications used by services are used both to aid donors in locating causes and projects of interest, and to give a sense of completeness – that most types of causes are represented through the service. From this latter point of view, it makes sense that the classifications used by services would share much in common with the current legal classification.²https://www.gov.uk/government/publications/charitable-purposes/charitable-purposes enumerates the legal categories defined in the Charities Act 2011.

For our service, there was an increased focus on completeness – a major part of the appeal being that we should reach a large number and broad range of charities, with the long-term ideal of “leaving nothing out” visible in our portfolio as early as possible.  In addition, because we would be asking our donors to engage actively with the classification, we needed to limit the overall number of categories, and with this in mind also decided against using subcategories.

We did not set a fundraising target at the start of the project – it was treated as an ongoing service, enforcing from the start the principle that donors did not select charities or projects directly.  While this created the problem of how best to keep donors involved, it removed the problem of how to get donors interacting with a large number of projects.

We made changes to our list of categories over the course of the project.  Details of the considerations behind the selection and changes are given in Appendix E.

Donor expression

Donors ranked categories of charity, such as health, children, and poverty.

Donors registered using a paper form, which presented them with the first list of categories above, and invited donors to rank the categories in relative order of importance using numbers from 1 to 9 (9 being the most important), and to leave a category blank if they did not want to support it.  The exercise was optional, and the form stated that leaving all categories blank would be taken as an instruction for the donations to be allocated at the discretion of the trustees.

Donors typically do not select all their charities at the same time and may never see the full picture of what they support, and while it is a positive move to make the overall effect more visible, we felt that it was asking a lot of donors, so we preferred to keep it simple by limiting the number of categories, just asking for a ranking, and not linking directly to budget implications.  We also worked on form designs asking donors to imagine allocating either the budget for the charity sector, or their own “lifetime” donation.

Work discovery

The Charity Commission’s database allowed us to find and contact relevant charities and projects, without relying on their active marketing.

Initially, we published the fact that we were seeking grant applications via our website and by writing to charities directly.  This was enough for other charities to find us, and produced a manageable stream of applications and requests for information.

A few years into the project, we replaced this method with a detailed search by our trustees of the Charity Commission’s database of registered charities, and we stopped accepting applications.  We give our reasons for this in the results section below.

With each approach, our overall goal was to produce a portfolio of projects.  We considered a project to be inappropriate for our portfolio if a significant proportion of people could be expected to object to any of their donations being put towards it.  This often rules out religious and political projects, for example, but not necessarily projects run by religious organisations.

When we moved to using the charity database, it was desirable to have a close map between donor expression and the classification used by the database, and the database classification matches the definitions found in charity law (a charity may claim its objects relate to one or more of the legal categories).  Instead of re-collecting donor preferences, we decided that our categorisation was close enough to these legal categories that it was acceptable to extrapolate the existing preference data to the new categories.

The Charity Commission periodically makes copies of their database available to download (see Appendix D), which allows anyone with technical resources to take advantage of common database features.  The features we used to aid discovery were filtering and random sampling.

Project selection

Taking each category in isolation, we discussed what sensible objectives for each should be, and chose projects through a multi-stage filtering and voting process.

The scope was treated as flexible for the majority of the project’s lifetime.  In particular, we had kept open the option of expanding the service to public users until we decided to wind up the service.  Delaying this decision had implications for whether our short-term focus should be on maintaining balance in the portfolio, or looking to emphasise diversity.  The final scope was to demonstrate a significant portfolio using only a small number of private donors making small donations over a period of several years.

When budget became available for a grant of reasonable size, the board would discuss charities on list produced during the discovery phase, until it was agreed that there were enough good candidates to choose from.  A shortlist would then be formed by each board member nominating two of the projects discussed, and a single project would be chosen through further discussion.

A grant offer would then be made to this project, stipulating that the grant was to be used for the day-to-day running of the project excluding any advertising and marketing costs.  The reason for this restriction is that neither of our discovery methods relies on our recipients advertising, and our problem statement makes strong reference to inefficiencies around advertising.  In the case of the first discovery method, each application was read and considered; in the second, any website containing sufficient information was considered.

Continue: how the service performed