Reimagining Sparrho (Part 2): What’s changed?

Continuing on from my previous post on the lessons I’ve learnt from Sparrho 3.0 as Head of Product, I’d now like to explain the major changes that we’ll be making in Sparrho 4.0.

Miriam (Mimi) Keshani, @MimiKeshani, is Head of Product at Sparrho.

The must: A straightforward solution for 95% of the problem

My analysis of the common pain points and motivations of our users showed that most would always check certain journals without fail as often as once a week to every day. I saw this same behaviour time and again, whether this was via manually visiting journals and their ahead-of-prints online, or relying on slow, and often incomplete RSS feeds, or even subscribing to 10+ email Table of Contents (eTOC) alerts, clogging up their inbox.

Within a bounded set of, say, 10 journals, you’d still be faced with hundreds of new articles every week.

Within a bounded set of, say, 10 journals, you’d still be faced with hundreds of new articles every week — trust me, I manually checked and gathered the latest articles from just 60 journals for the first nine weeks of building Sparrho 4.0, and it was certainly painful! Sparrho versions 1.0 to 3.0 excelled in helping our users find content that they didn’t even realise they were looking for, but what I learnt was that there was already too much to read from where our users were already purposefully looking.

This meant Sparrho was becoming just another place to feel guilty about not checking, and not providing a solution to 95% of our users’ struggle with information overload. Our next step? We decided to test a journal aggregator.

The goal: A service you can’t live without

Sparrho Journals beta platform, November 2015

We knew our solution had to be at least 10x better than any existing journal aggregator. None of the currently available tools have been specifically designed for scientists, so we homed in on the needs of PhD students and postdocs. We created a beta platform (pictured) to heavily test with a select number of chemists at top UK universities, and we set our aim to be as accurate as the original journal website for these chemists’ favourite journals. We’ve worked hard to improve our mobile compatibility and browsing experience too.

Journal feeds designed for your convenience

We’ve found journal RSS feeds to be notoriously unreliable.

Throughout the past years, we’ve found journal RSS feeds to be notoriously unreliable, often selectively missing entire articles, and key information like graphical abstracts and figures. However, RSS feeds do allow you to mark each article read individually, rather than have to deal with the long and poorly formatted textual lists of articles in journal eTOCs.

We wanted to make browsing as easy and contextual as possible, so we decided to allow users to group journals into custom ‘channels’, which could, for example, be categorised by the groups/collaborations they were working in, or by specific research themes — and not the categories assigned by publishers. Sparrho 4.0 will always give graphical abstracts and figures directly in the feed where available, unlike most journal RSS feeds.

Sparrho is currently the only platform that presents research articles in a consistent format regardless of the original journal or publisher.

In essence, Sparrho is currently the only platform that presents research articles in a consistent format regardless of the original journal or publisher, and intuitively allows users to mark individual items ‘as seen’. When dealing with different publishers’ formats, this really isn’t a trivial task. Our tech now provides an easy browsing experience that researchers have long deserved.

Articles stored for your easy access

The mindset and motivations of a user differ vastly between finding articles and reading them.

My behavioural studies showed that the mindset and motivations of a user differ vastly between finding articles and reading them, the two activities often occurring at different times of day, or on different days entirely. Therefore, it was clear that we needed a place where articles could be effortlessly saved for later reading, organised, and shared. This led to the creation of ‘pinboards’.

Email is still the most common form of sharing and recording ideas between groups and colleagues, but comments, and often even the article itself, get lost along the way. Pinning articles to pinboards can keep the entire conversation in one place. This is perfect for journal clubs where each meeting’s list can be stored, and the discussion remains visible to everyone long after the event. We’ve even had a pair of international collaborators, based between California and London, who use a Sparrho pinboard to keep track of the latest content that could be important for a review they are writing together.

The (temporary) compromise: Old and new in parallel

We’ve decided to keep the old version (3.0) running in parallel on old.sparrho.com, as most people will also need to filter their channels by keywords too.

Today we will transition sparrho.com to version 4.0. However, we’ve decided to keep the old version (3.0) running in parallel on old.sparrho.com. There is a good reason for this: a key feature is still missing from this new version –keyword filtering and search. We know that just looking at certain journals is not enough to solve 100% of the problem; most people will also need to filter their channels by keywords too. However, we need to seriously overhaul the tech behind this, so old.sparrho.com will keep running until keyword filtering can be fully incorporated into the new version.

I think it’s important to explain why some seemingly key features are currently missing from Sparrho 4.0:

  • Search and keyword filtering — coming soon — we’ve learnt a lot about what academics need from good search and recommendation algorithms in the past few years, but as my last blog explained, version 3.0 was trying to cater to too many different people. I’m excited to streamline this knowledge into a single place for academics to create highly customised and specific channels for their needs.
  • Email notifications for updates in channels — daily or weekly alerts coming soon, watch this space!
  • Recommendations — we’ve found that currently the highest quality recommendations we can give come from people, and not machines (yet). Sparrho 3.0 required too much investment to provide continually high quality recommendations, so we created pinboards to allow our users, experts in their fields, to share and discuss ideas. Having said that, creating ‘smart’ channels that learn is something that we’re constantly testing and will be slowly re-integrate as we work out the kinks.

Rolling out, with patience, discipline, and resilience

One thing we’ve learnt and brought in version 4.0 is the need to build quality tech that is resilient. In the words of Viktoras Jucikas, Founder of YPlan:

“The goal isn’t to solve all possible future problems now, but rather to not carelessly do things that make those future problems unnecessarily harder. Therefore our systems are built in such a way that they can be changed, and improved.”

For instance, our content database has been totally re-designed so all items are sensibly structured and it works extremely fast. (You’ll be able to learn more about the tech behind Sparrho in future blog posts.) However, building resilient tech also means we can’t build everything we want at once, and so prioritising has been painful, but necessary.

We’ve learnt what we can and we’re marching on

Startups are about solving problems, not creating products, and sometimes they need to change to grow. Rather than assuming we knew what people wanted, we listened to the our users’ underlying problems, and worked to solve them, even when they didn’t fit the current iteration of our product, but still fit in our wider vision to improve accessibility to cutting-edge research. We’ve learnt a lot and also built some amazing tech too — Greg, our Head of Engineering, will be writing a blog post to fill you in soon.

I’m excited for you to try Sparrho 4.0 and would love to hear your thoughts, feedback, or ideas. I’ve given a highlight of our plans but what other changes do you think would help better incorporate Sparrho into your literature discovery process? As you know we only plan to grow in the most useful way for you, our users. Thanks for helping us get this far!

p.s. We’re always open to recommendations for new sources to add to Sparrho, so please drop Mimi an email at mimi@sparrho.com to tell us what you’d like to see next.


Originally published at blog.sparrho.com.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.