The Big Read, sponsored by Vodafone: Having recently written about Dublins unique advantage in Big Data and its potential to become a global hub for this flourishing sector, the next logical step was to get a first-hand account from one of the companies contributing to the citys growing prestige in data analytics.
Michael Kavanagh (pictured above, centre) is CEO and Co-Founder of Social Honey, an audience platform that helps publishers make money by understanding their audience in real time. I caught up with him recently to find out what makes Social Honey different. Somewhat unsurprisingly, given that he makes his living providing actionable insights to publishers, I learned a lot more than I bargained for, from the importance of truly understanding your customer and how to create genuine value for that customer, to the resilience required to get a startup off the ground, and what it is that sets Dublins data scientists apart.
About Social Honey:
Kavanagh speaks in the very deliberate, precise manner characteristic of many scientists when describing the Social Honey platform and what it provides to customers. Those customers are publishers, more specifically big media brands with the volume of content and readers that produce the extremely large data sets with which Kavanagh and his team of Data Science and Machine Learning PhDs work. At its simplest, Kavanagh explains, Social Honey gives readers what they want to read from publishers and allows publishers to engage on a per-reader basis, in a personalised way, that gives them more value from their product.
Social Honey achieves this by first building a base audience platform with anonymous profiles of a given publishers readers, leveraging Natural Language Processing (NLP) to serve up key information to the publisher about specific readers landing on their site. The software uses a scoring system of probability to create personas and categorise readers based on their interests: for example, Kavanagh says, a reader might be 90% a female, 70% a fashion lover, and 60% interested in arts and entertainment news. Social Honey also refers to third-party sources like Google Analytics, Facebook, and Twitter to validate its user data: it then uses this data to provide the reader with more relevant, valuable content, presenting toast notifications to recommend further articles that are likely to be of interest within the site.
The publisher is presented with a real-time, dashboard view of who its audience is, organised by article and author, what its audience is engaging with and for how long, and which content is winning. Publishers can also see which channels they are acquiring readers from, i.e. direct traffic to their site, search engine results, social media, and internal traffic from other articles on their own site: the latter being the metric which Social Honey aims to maximise by ensuring that the most relevant and interesting content is placed in front of the reader, at the right time, in an unobtrusive way. Kavanagh and his team have done this so successfully that click-through rates to recommended articles increased in some cases by 1000%.
So: how did they achieve such a marked difference, when compared to standard, off-the-shelf related posts applications like the aptly named Yet Another Related Posts Plugin (YARPP)?
Knowing your customer (and your customers customer):
The answer to this question starts with gaining an in-depth understanding of your intended customer; it became immediately apparent to me upon meeting Michael Kavanagh just how important this was to Social Honey. It didnt take an investigative journalist to figure that one out, as the Social Honey team are currently located in the Dublin offices of major publisher (and Social Honey customer) DMG Media, an arrangement which arose from a successful working relationship, the benefit DMG saw in cohabitating with a vibrant start-up, and Social Honeys eagerness to get under the skin of the publishing industry.
Social Honey didnt always know their customer as well as they do now, as Kavanagh freely admits: Publishers said ‘We want recommended content’. We said sure. We proceeded to come up with a bunch of different terrible approaches – really bad approaches.” Having identified the need among publishers to keep readers on their site and engaging with more of their content for longer periods of time, Social Honeys first attempt at solving the problem using an active reader feedback (1-5 rating) system failed. What Kavanagh learned from this experiment was that the publishers customers – i.e. readers of online content – conform to a 1-9-90 rule with 1% being classed as creators, 9% engagers, and 90% passive consumers who would inevitably skip the step of rating content online.
Since then, the Social Honey team have dedicated themselves to understanding publishers and the world in which they operate. “Publishers are now in a tech industry so they have to think about things differently”, says Kavanagh. As CEO of a SaaS company, he is acutely aware of concepts like customer acquisition cost (CAC) and life-time Value (LTV) and the impact these metrics have on the success of his business. The SaaS industry has influencers like Jason Lemkin and Tomasz Tunguz to ensure that its founders are intimately familiar with such metrics, but online publishers often lack that same understanding of their metrics. Social Honey, for its part, aims to bring this insight to publishers, allowing them to see each reader as a customer and make better informed business and editorial decisions: “In order to maximise return and calculate the CAC and LTV of a reader,” says Kavanagh, those metrics have to be worked out and, in our opinion, that metric can only be worked out by understanding what the value of the piece is to the reader”.
Creating genuine, measurable customer value:
Social Honeys steadfast customer focus on solving fundamental problems for the publishing sector is set apart by the integrity with which it is going about these tasks. Kavanagh remains averse to over-hyping the companys achievements and is unwaveringly matter-of-fact in his description of what they have delivered. “We concentrate quite specifically on statistical significance in everything we do.
Data is at the heart of everything Social Honey does and, in data, if youre doing it right, everything boils back to statistical significance. Michael Kavanagh knows that the value he delivers for his customers evaporates if the data is proven to be flawed: “What’s the point in having a bunch of guys with PhDs in Machine Learning if you’re not going to do some math around it? To not just kind of think that something works, but to mathematically know that it’s completely sound is much better.
As far as Social Honey are concerned, when the reader is provided with more valuable, relevant content, the publisher stands to make more money, be that in the form of serving the publishers display ads, or a more direct subscription-based transaction: “In comparison to traditional media, we see an entirely different view of what’s happening when a consumer hits a website. We see a consumer spending money in the form of their time.”
Helping publishers to understand their readers by providing them with actionable data in real time, and helping readers navigate through the noise to find the content they are searching for are propositions that carry a lot of weight. Not only that, but Social Honeys data-centric approach ensures that the benefit gained by its customers is measurable and the ROI is clear.
The Dublin Factor:
Having decided to move himself and Social Honey back to Dublin the week of Web Summit 2014, I was interested to find out if Kavanagh agreed with my earlier hypothesis about Dublins advantage in Big Data. It turned out that he had a unique take on the origins of the Irish capitals status in the area: “Because so many multinationals have based themselves in Ireland over the years, so many Irish graduates have been exposed to much larger infrastructure support and architecture at a very early stage in their careers. That coupled with a highly-skilled workforce in machine learning, a lot of people with mathematics degrees coming out, come together to form a recognition of what scale looks like and an appetite to play in it.”
As with the founding team at Boxever, the Social Honey founders come from the airline industry and have been part of teams that have operated at massive scale: Kavanagh cites the team of four that ran the booking process for Ryanair.com, Europes largest travel website. After that, scaling out huge architectures to accommodate vast amounts of data didnt seem so big or so difficult, and this is a factor that he believes is common to a lot of Dublins successful big data start-ups: they have gained invaluable experience of huge data sets working at the various multinationals located in Dublin and have carried those skills into their own ventures.
Add to that the distinctly Irish resilience that keeps these startups moving forward and trying new approaches (regardless of the amount of failed attempts suffered) and a clearer picture of their propensity to succeed emerges. Social Honey appears to have all the characteristics of a future SaaS superstar and another success story for Dublins thriving startup scene.