Sparrho (www.sparrho.com) is a discovery tool for researchers. The user registers their interests and Sparrho provides them with "feeds" (largely abstracts, cut can be videos or papers) of relevant content. So clearly Sparrho appears to meet a need.
They claim it replaces Electronic Table of Contents (ETOC) alerts provided by publishers for new issues of journals – but these were moribund already. How does it work? Firstly, Sparrho searches lots of journals – just 21,000 of them. The user then creates create “channels”, subject areas of interest to the user, such as “Wikipedia” or “ash dieback”. And then Sparrho provides details of new content in this area. In addition, users can copy abstracts to a shared area called a "pinboard". Articles held on the pinboard can be commented on by the user. The comments (and the pinboard itself) can be private or shared with anyone else the user has given access to.This idea is certainly not original – many services are competing to capture the shared notes idea for academic users.
What are the limitations of Sparrho?
- Sparrho has a very journal-centred view of the world. When you set up your channel, you are asked first to “filter by journal”. You don’t have to do this, but it suggests that your searching will be by journal primarily, rather than by term.
- Sparrho tracks many types of information, but for journals, which is one area where researchers expect complete coverage, it doesn’t track them all. If you want to add any more journals, you have to add an RSS feed yourself. There were already 28,000 journals in 2012, according to the STM Report, so I would expect there to be at least 30,000 today – in other words, Sparrho is currently only tracking two thirds of the scholarly literature (measured by journal)
- You have to trust what Sparrho is doing behind the scenes – you have no way of checking and they aren’t about to explain their methods. There was a reference in early posts about Sparrho that there would be buttons to “like” or “unlike” content, and these preferences would them be used to improve article selection, but I see no trace of a “like” button on the current site.
- Sparrho looks to be a very small organisation. A photo illustrating their growth plans shows a table with just six people around it. They talk about receiving “six-figure” funding, which would be enough to pay for a small number of people for a year or two – not beyond.
- Founded by a biochemist, Sparrho appears to be focused on science rather than humanities and social sciences.
An interesting additional feature of Sparrho is “perspectives”, which it describes as “two-minute research digests”. These perspectives are an interesting, if slightly odd, feature. The aim is sensible: to make articles more intelligible. So each research perspective restates an article abstract in a more user-friendly style. The perspectives on the site have been written for the purpose, and the website offers money to researchers to compile more. But to do this on a large scale would take vast amounts of time and money (there are 6,000 new scholarly articles published each day), so it’s not surprising that the examples on the site are sponsored by third parties. There is no sign that Sparrho has a viable model for this activity, and I don’t think it forms an integral part of what Sparrho does; if it was quietly dropped in a future release I wouldn’t be surprised.
How good is Sparrho? One problem is that there is a lot of competition in this area. Sparrho was featured in one conference seminar in 2015 as just one of seven startups, each given five minutes to present. Heavyweight competitors include Google Scholar’s own recommendation engine, and an increasing number of recommendation services provided on publishers’ own websites. Yet the problem for researchers is a very real one. One review in Nature News compares Sparrho to a simple trick a researcher came up with, a utility to trawl PubMed servers for just one term, “Drosophila”, and email him the details. For some subjects, that is all you need. Can Sparrho beat that? To answer that question will take some time while researchers find out what it can discover for them – the key question will be if it finds content they would not otherwise have found.