How publishers can combine predictive analytics with CRM to convert registered users into paying subscribers

I know I’m wasting half my marketing budget, I just don’t know which half

It’s a well-worn marketing phrase and reflects a problem many would like to solve.  The better you understand your customers, the more effective your marketing – so we’ve been looking at how to apply the profiling technology we use in advertising to CRM, to see if we could predict which ‘unknown’ customers are most likely to buy.

VisualDNA has invested heavily in personality profiling technology. Using patented visual-based personality quizzes, we are able to profile individuals online with great detail and accuracy, and subsequently use behavioural models to extrapolate this detailed knowledge onto hundreds of millions of users. We group people by various characteristics into ‘segments’ – anonymised groups that can be used by businesses in communications, usually targeted advertising, product recommendation or content personalisation.

There are other applications too, one such area being CRM. Most businesses, whether they are business-to-business, or business-to-consumer, possess a database of customers, or active leads. And those customers can be categorised in all sorts of ways. But whichever way you define your customer base, it’s notoriously hard to predict who is most likely to buy.

So we wanted to test the idea that the inference algorithm we use to build our high-performing segments could also be used to identify which customers are most likely to buy out of any given data set.

Piano Media is a company who run a subscription paywall business on behalf of a large number of publishers in multiple countries in Europe; with a massive data set of customers who fit one of three categories;

1) Unknown / un-registered,

2) Registered but not subscribed,

3) Paying subscribers.

Clearly the objective for any business operating a freemium model is, having attracted the customer in the first place, migrating your unknown and registered users into paying customers. The question is, how do you know which of your registered users are most likely to become paying subscribers? At which part of your audience do you direct your marketing?

Working with Piano, we drew a random sample of 30,000 registered non-subscribers and compared their online behaviour to that of the entire population of paying subscribers; using what we call Look-alike modelling. We did this completely anonymously, using randomly generated anonymous identifiers stored in cookies.

From here we were able to assign a similarity score to every one of the 30,000 registered users – the score representing how similar user’s behaviour is to the paying subscribers and therefore how likely we think this person is to convert to a paying subscriber.

Piano then did three mailshots, under exactly the same conditions, to three separate groups within that population

• The top 5,000 (those with highest score)

• A random sample of 5,000 users

• The bottom 5,000 (those with lowest score)

The results are hugely encouraging. On the top 5,000 Piano got a conversion rate of 1.59%, compared to 0.52% for the random group and 0.22% for the bottom group. In other words, our algorithm correctly predicted those most likely to convert, based on the theory that the online behaviour of a paid-up subscriber was sufficient to build a profile on them and use that as a means to find those with a similar profile. Moreover, we correctly predicted that users with behaviour dissimilar to paying customers are a lot less likely to convert than random users, so that marketers don’t need to spend the money and energy marketing to these.

This result has significant implications. It means we can, in theory, take any database, whether online or offline, anonymously profile anyone in that database, and use our inference and Look-alike algorithms to predict behaviour to improve the effectiveness of a company’s marketing – in this case increasing conversion more than three-fold vs a random sample and weeding out those least-likely to buy.

Find out more.

How publishers get a real time understanding of their audience through WHYanalytics

This interview first appeared in US digital marketing blog The Makegood under the title “VisualDNA’s Ian Woolley on Real Time Understanding”

Ian Woolley is Chief Commercial Officer at VisualDNA, an audience insights company with patented technology helping businesses understand people online. The company is headquartered in Shoreditch, London and Ian joined the team just a few months ago. The Makegood recently spoke with Ian about VisualDNA’s products and vision.

The Makegood: Congratulations on the launch of Real Time Understanding (RTU). Could you explain how this technology works to profile visitors based off their emotions?

It starts with our patented visual personality quizzes. From these, we can create incredibly rich and accurate personality profiles very quickly.  They’re rich because the combination of images a person picks can be far more insightful than a list of options and require less interpretation than a multiple-choice of words, and they’re accurate because people are incentivized only by wanting to know more about themselves – we find this makes the answers genuine.

From here we use statistical analysis to put people into clearly-defined groups and behavioral inference to scale those groups into what we call Emotive Segments – anonymized groups used for targeted advertising.

Real Time Understanding profiles a website’s audience in real time, is viewed through the WHYanalytics platform and gives the publisher a deeper understanding of their users by demographic, intent, brand preference and personality type.

The Makegood: This is a very novel idea for marketing. What inspired you to create this technology and launch VisualDNA? How do you see it impacting digital advertising?

It was the other way round, really. Our founder Alex Willcock started VisualDNA because he saw the potential in using visual quizzes to understand people and improve their experience of the internet. If people have the means to communicate their intentions online in a way that’s useful for businesses then the benefits to both parties are clear.

Since then we’ve built a successful business around targeted advertising, and see a huge opportunity in emotive segments. It’s a nascent market at the moment but two big trends are in our favor; first, we know brands already use emotive characteristics to define their customers – the appetite is there to move beyond demographics in targeted advertising and second; the wider market for data is growing as programmatic ad buying platforms get more sophisticated and make it easier to use, publishers are recognizing that programmatic is not a threat to their existing revenues and more brand campaigns are becoming automated.

This all adds up to a bigger data market that we’re ready to capitalize on.

The Makegood: “Big data” is a buzzword today. I see you are also launching WHYanalytics. Can you explain what type of data will come from this and how it will benefit both marketers and content creators.

WHYanalytics essentially visualizes a publisher’s audience using our Segments, thereby revealing personality traits with Emotive Segments – as well as demographic, intent and brand preferences.

We’ll analyze their traffic, benchmark it against our database, and play that back to them – all in real time. To choose a segment at random, in any given online population you’d expect a certain proportion of “Creative Contemplators”. We will assess how that segment typically behaves online. We then look at the proportion of users on the publisher’s site exhibiting those characteristics – that’s where the benchmarking takes place. If the size of that group is well over average (say four times as large as you’d expect) then WHYanalytics will tell you that.

The upshot is that you as a publisher now know that your audience comprises a particularly high number of  “Creative Contemplators” than the norm – valuable insight when deciding what to put on that page, whether it’s an ad, product or editorial.

The Makegood: Many users express concern with privacy invasion when it comes to content personalization. How do you see companies such as Visual DNA handling this as Artificial Intelligence programs advance?

Our approach is to be transparent about what data we collect and how we use it, and we look to create value for people in the process. It’s worth emphasizing that the data we collect stays with us and is not sold to anyone else: the segments we create are completely anonymized.

On the wider issue of data ownership, we think people would be better served if they owned their own data and could use it in a way that creates value for them.

Making that happen remains a long-term goal for VisualDNA but neither the technology or the market is ready just yet.

The Makegood: Real Time Understanding seems to be a breakthrough technology with a lot of potential. What challenges do you see it facing as it progresses?

In the short-term our challenge is to help the market understand the benefits of data, and specifically the segments that we create out of it. Online advertising is a complicated industry and sometimes even those involved with using data on a day-to-day basis struggle to use it effectively – and on top of that the market is moving so fast.

So the first step for us is making the market aware that there’s now high-quality Emotive Segments available, and helping those in the market for data to use it to make their campaigns more effective. WHYanayltics is a great way to visualize an audience in terms of our segmentation, and we have plans to add new features in coming months that will make that information more actionable and measurable.

But we know that everyone in the entire marketing value chain wants to understand people better online – the building blocks are now in place and we’re really excited about the potential.

The Makegood: Thank you, Ian.

See more:



How to use data to optimise targeted advertising campaigns

This article originally appeared on The Drum on 13 September.

The growth of big data has seen a rapid increase in the use of data segments to power targeted advertising, the science of reaching anonymised niche groups of people defined by common characteristics.

VisualDNA uses the first-party, self-declared data we generate from visual personality quizzes to create high-quality Demographic, Brand Preference, Purchase Intent and, now, Emotive segments that power targeted advertising.

With multiple factors affecting the performance of a campaign much like effective media buying effective data buying is something of an art form. Programmatic? Not really: data optimisation is done manually, and by smart people. Our top tips for using data segments successfully:

1. Beware of ad-blindness! Don’t over-impress…

A car brand running a campaign of 50 million impressions will likely want to reach people in-market for a car. In theory, a segment such as “Car Buyers” would yield great results, but if that segment comprises just 500,000 people then each user would see 100 impressions a month. Not good.

Over-targeting an audience that’s too small bombards your audience – reducing conversions and increasing CPAs per impression. So define the total number of impressions served and work back. If an average of 15 monthly impressions per user is about optimal on a cost-per-impression/ conversion basis, then a 500,000-strong segment is only enough for a reach of 7.5m impressions.

A run-of-network to make up the remaining 42.5m impressions means you’re reaching a wider audience, reducing the average data and media cost per impression –plus you can test the performance of the data. Data can dramatically increase conversions, but you won’t hit that uplift and justify your data spend if you over-impress.

Trading desks know this of course but thinking about data at the media planning stage helps them buy better segments. The IAB’s Data Usage and Control Primer is a great intro for anyone using data as part of a campaign.

2. Use analytics tools to define and build your audience

At face value, choosing segments should be straightforward.  Quality data, demographics from trusted providers such as Experian, and intent data from sector specialists should perform. Descriptive VisualDNA segments such as “iOS Preferrer”, or Emotive segments, built from self-declared data collected through our quizzes, are also obvious ways to build an audience.

But not all segments are self-evident. Audi, for example, might intuitively buy an “Audi Preferrer” segment from one provider, but it might be that a “Luxury Holiday Preferrer” segment from another provider is a better fit for their brand.  Adding such a segment to a campaign would increase the size of the targeted audience, improving the performance of a campaign

In addition to industry-standard audience insight tools such as Quantcast, WHYanalytics profiles a website’s audience to show which VisualDNA segments rank most highly – powerful new insight for publishers, ecommerce, advertisers, media agencies and brands.

Such tools help publishers meet high-value brand campaign briefs or identify niche audiences while, in ecommerce, online stores can personalise around personality. Plus agencies can pick the best segments to build a targeted brand advertising campaign.

3. Set the right metrics, test the data & monitor throughout.

What does success look like? A Victoria’s Secret creative showing an attractive woman sat on a beach may attract a disproportionally high number of clicks from men, suggesting that male-skewed segments should be used, at the expense of other female-oriented segments…

So it’s about conversions above clicks but track both with a short, un-targeted, “Discovery Phase” at the start of the campaign against each creative.  Whether it’s a direct response or brand campaign, by analysing the ad (impression beacon) and conversion page (conversion pixel) we can tell you which segments are responding. From here you can optimise around the right segments, and have a benchmark from which to measure performance.

We’ve created WHYanalytics to help publishers, ecommerce websites, advertisers, brands and agencies better understand VisualDNA data, and use it to create a better-personalised experience for their audience – we’re working with our beta partners to build more features and functionality that will make it easy to track. Free to use and easy to deploy, try it out at

Why the future of digital marketing is understanding the person – not targeting the cookie

This article originally appeared in Real Business on 16 September

Big data is changing how businesses speak to customers. The internet has made the world smaller, more connected, but it’s made your potential market larger, more complicated, and harder to understand. Big data is stepping in to fill the void.

But we’re still trying to fill a digital hole with an analogue peg.  It’s understandable that marketers might want to treat online as a digital extension of the physical world.  Media buyers (the clue is in the title) buy media, and even those who now talk of ‘buying audiences’ – usually define audiences by the offline demographic terms that have served us since the Mad Men era. And as long as TV remains king that’s not going to change. Even in digital the majority of spend is direct, through a publisher or ad network, because that’s what we know.

Programmatic buying & real-time bidding is changing this. Instead of buying a publisher’s audience, you target a user, wherever they may be. What started out as a technical means to sell ‘remnant inventory’ is now, for some, the primary mechanism for ad sales.

Publishers have mixed views on this. Some are terrified by the prospect of eroding the value of direct sales, while others embrace it as a means to trim sales teams.  Whatever the view, it’s happening, and if the person, rather than the platform, is your ‘target’ it follows that the better you understand that person, the better chance you have of connecting with them.

This narrative isn’t just relevant in the narrow field of online adverting. It has profound implications for the way businesses and people communicate. Taking a purchasing decision as simple as buying a chicken, for example, there’s an abundance of choice – corn -fed, free-range, economy, pre-basted etc. Demographics can guess at what media ‘those people’ read, and analyse what prompts ‘those people’ to choose one above the other, but such purchasing decisions are more closely linked to personality – what kind of person they are.

But understanding personality is incredibly difficult. Brands are remarkably sophisticated when they profile, and mass media advertising can send out signals in the hope that they are paying attention, but how does that work in a modern digital ecosystem when we have to go to them? Who is the person behind the cookie?

VisualDNA’s patented personality profiling technology uses data from visual quizzes to create an online persona. And from these, we use an inference algorithm to scale particular traits into anonymised segments that are used in targeted advertising.  The market for data in advertising is still dominated by demographics, but things are changing.

Having witnessed a shift from targeting the media people consume, to targeting the person direct (as defined by the data created through cookies), big data is helping us understand the person behind the cookie. But we still have a long way to go.

Products like WHYanalytics, powered by VisualDNA’s profiling technology, can help marketers better understand their customers’ psychological profile. By making that information useful for marketers, in real time, and in a way that’s relevant and valuable to the consumer – we can focus more on understanding the person, and less on targeting the cookie.

Infographic – how VisualDNA creates emotive data from quizzes


Is emotive targeting the future of digital advertising for both publishers and brands?

This article originally appeared in MediaTel under the title “It’s Time To Get Emotional“.

By Ian Woolley, Chief Commercial Officer of VisualDNA

Innovation in advertising technology means it is now technically possible to put a specific ad in a specific place at a specific time – for a specific price. Coupled with this, the explosion of big data means our understanding of the person behind the screen is getting better and better. Programmatic media buying brings this powerful technology and rich data together in one place.

In that light, it’s not surprising that programmatic sales are growing, with the latest IDC report* stating that 52% of UK digital inventory is now traded automatically. So why is it that so much online advertising value remains in direct brand campaigns – that is, where advertisers buy impressions direct from publishers?

It’s all about data. For all the technology in the world, an advert can still miss the mark if the advertiser has little real understanding of the person viewing the advert, or what they actually want.

Simply put, advertisers trust that publishers understand their own audiences. In other words, they have great data, and are therefore by extension the most capable channel through which to run targeted campaigns. Further to this, some inventory can only be bought direct as some publishers continue to withhold premium inventory from programmatic platforms, concerned that it may be devalued.

So while third-party data has the potential to power super-targeted, relevant advertising, neither publisher or advertiser is currently capitalizing on the technology or data that’s available.

Is there a way for how can brands to optimise marketing spend with accurate data-led targeting, whilst allowing publishers to maintain the value of their premium inventory? Perhaps the answer is surfacing in the form of emotive data.

Moving beyond the limiting demographics of age, sex and income, emotive data reflects a person’s personality, motivations, outlook, interests and attitudes. Our method is to collect this data through opt-in personality quizzes, creating rich profiles, group people with similar traits into anonymised ’emotive segments’, to power more relevant advertising.

We anticipate a growing demand for emotive data from both advertisers and publishers. Advertisers recognise the value of the right message in front of the right person in the right context – not just because it’s a waste of time and money to talk at the wrong person – but because of the potential damage to brand reputation when things go wrong, as illustrated when Sky and M&S pulled ads from Facebook in late June.

In turn publishers need to prove to advertisers that their customers can be found, and reached, in their audience, and raise the value of their inventory. The inclusion of emotive data in inventory offerings can provide that proof and increase the value of ad sales; whether it’s powering targeted high-CPM brand campaigns through ad networks or non-premium inventory through programmatic.

We all know brands want to connect with people on a deeper more human level, and there is huge potential in emotive data to bring together brands’ customers and publishers’ audiences using personality traits. With the means to move beyond basic age and gender targeting, it’s getting easier for advertisers and publishers to connect with an audience on the basis of who they are not what they are.

Do Brits and Americans conform to type? VisualDNA in The Times & Daily Mail

In a sample of 20,000 VisualDNA quizzed users (10,000 from the UK and 10,000 from the US) we looked for answers to the plagued questions of national stereotypes. Are Americans more optimistic? And do Britons actually have a stiff upper lip?

See our coverage in The Times below, and on Mail Online.


Emotive Audience Segmentation: The Science Bit

At VisualDNA we’ve developed a patented way of uncovering a person’s needs, their desires,  and the things that are important to them.

We start by promoting free visual personality and lifestyle quizzes that give users insightful and astonishingly accurate feedback about themselves.

This is especially true with our newest, most advanced quiz, Who am I?, which classifies people according to the Big 5 factor theory of personality: openness, conscientiousness, extraversion, agreeableness and neuroticism.

The process of turning the results of a visual quiz into a rich, accurate online audience profile isn’t completely seamless or straightforward.

Making sense of terabytes of quiz data and scaling the solution out to a global audience network is as complex as the personalities of the people who take our quizzes.

In fact it’s far too complex to go into in a blog post. So, we asked our team of data scientists, creative developers and psychologists to sit down and tell us how they do it, and put it all in a white paper.

  • Why do we use image rather than text-based quizzes?

  • How do we extract meaning and understanding from the full response pattern of a user?

  • What kind of algorithms and mathematical formulae do we use? (the real science bit)

If you’re asking yourself any of these questions, then we think we’ve got the answers for you.