Music Ally

Today we publish the second in our five part series on how internet intelligence can be used to create new metrics for measuring music. Yesterday we spoke to Eric Garland of BigChampagne; today we interview Alan Ault, managing director of WaveMetrix, which describes itself as “the global leader in buzz research”.

Alan’s comments can be found as part of an extensive feature analysing the market for music metrics, alongside observations from other leaders in the field including executives from MusicMetric and Nielsen BuzzMetrics. To read the feature, log in or sign up for a free trial of the Music Ally Report.

Continue reading after the jump for some enlightening example output from WaveMetrix research as well as Alan Ault’s  explanation of why listening to the public using internet data can point brands in the right direction even when the media are barking up the wrong tree.

How did you get started in this area?

We’ve been working since 2003 doing buzz research. It started as an offshoot from a consulting company.

There are a number of different challenges to uses online buzz. First is collecting data – there’s lots of stuff in all sorts of different places and the challenge has grown with the internet, whether it’s YouTube or Twitter or the Times Online – so we built our own platform that helps us collect discussion.

What makes your system different to the competitors?

We spent a long time working with a number of products to use artificial intelligence to understand discussion. Most tools do that, looking at the post and generalise about it – for example, are comments about Obama positive or negative? But this doesn’t work on multiple languages, colloquialisms and complete consistency isn’t there. I want to know the way my artist is being talked about. When they talk about an Easter egg on a DVD or a live event, I want that categorised under event / Glastonbury / 2008 / and I need it always to be the same, not sometimes to be somewhere else.

The only way to do it properly is to use human analysts is to code discussion. This is the way it’s done in focus groups, where consumer research uses humans to code research. Our process brings lots of analysts to bear to read and code in more than 35 languages from anywhere in Europe, China, Korea, Philippines, Indonesia, the Arabic countries and more.

Another key difference is understanding what drives sales. So rather than saying ‘the most talked about thing is how they look’, that’s interesting but the crucial thing is why does someone recommend it? If I’m talking about the new album from Them Crooked Vultures, I might talk about their history, the fact that they’re a super group, but I might say the reason you should buy it is because of track five. They might say Spotify is great because it’s free, or because if it’s got a great range of songs – these are different things.

The recommendation is subtly different from the rest of the discussion.

How do your relationships with clients work?

The way it starts is an ad-hoc fashion  - for example we work with Sony BMG and they’d typically call us up and say we want to do something with artist A or Nokia might say we’ve got a new variant of Comes With Music launching next month and we’d like to do something with it

We usually start in a small area, leaning about something you can action in a tactical way and from there start larger monitoring programmes that cover a wider range

If you’re doing an album you can ask what’s the buzz from hardcore fans versus more generic members of the public and what’s the buzz from lapsed fans – people saying ‘I used to be a fan but I stopped buying because I got fed up with x’ – what do they think about it because they should be an easy win.

What kinds of data sources do you cover?

We would cover any place where there’s any consumer comment – you can’t say ‘I’m not going to cover that area just because it’s technologically difficult to do’ – each category tends to have discussion grouped in different places – in consumer electronics there are lots of blogs, lots of forums dedicated to specific elements of consumer electronics for example TV. There are lots of reviews and reviews are very different to discussion, plus you need to be careful not to confuse press and consumer reviews.

Why is it so important to differentiate press and consumer comment?

People in the press feed off each other…at the time that the Sony PSP and the Nintendo DS launched, everyone thought the PSP was the best and the journalists were feeding off each other. We did some research at the time that suggested the DS would be more popular with consumers, which seemed counter-intuitive but ultimately the DS did turn out to be more popular. Comments on pro blogs or news articles are fair game but as for the actual blogger or semi pro blogger we don’t count them. You wouldn’t invite Jeremy Clarkson to a focus group.

Can’t you get a lot of this data just by using free tools on the web?

The answer is yes you can – and for some people that’s totally appropriate. There are plenty of free tools and that’s a great place to start – if you want a couple of qualitative comments that’s fine. It’s pointless asking us to do something like that and all companies need to be doing something like that themselves anyway. The first thing we do when we sit down with a client from a research perspective is ask ‘what is it that’s keeping you up at night? What don’t you understand about Spotify or Last.fm?’ From there we’ll answer their research objectives.

We can do it consistently in ten countries, that’s going to be difficult to do with work experience people and free tools. We’re about delivering powerful insights so as we work with a lot FMCG companies where they have people working viral programmes…we can do research and measurement to measure the success of their viral campaigns. If you want someone independent to measure the effectiveness of the marketing being done by another agency we can do that.

What kind of metrics do you take note of?

Online buzz isn’t just about what’s being said – and it’s often not what’s being said at all. It can be what’s being read? If we’re on Digital Spy and I say ‘Did you see X Factor’ and then I get into a discussion about Strictly Come Dancing then was that thread about X Factor? How do I count that in terms of mentions?

We’re based on each individual opinion – and we can capture how many page views there are of an opinion. We measure buzz on pageviews of each opinion

Bette than mentions because it tells me what are the most important sites in terms of readership and that’s a key measure

Then let’s get subtle – the next key measure is net advocacy – how many pageviews of recommendations for my brand versus against my brand .

We’re just finishing a piece of work using specifically our advocacy measurement and we’ve done regression analysis on film takings because its easy to get the data – really strong correlation – really well and much better than just measuring buzz

How much of your information is “live” or “realtime”?

We can do turnarounds very quickly – we can do daily reports and have a portal that can deliver the raw data more quickly than that – but it depends what you want to do – if there’s an issue you want to know about it quickly but to understand the impact you have to know if people are commenting about it over time.

One person posting about a bug doesn’t tell you much – you have to see if someone else says that there’s a bug. You have to know if people are reading it

The issue is less to do with our technology or monitoring social media and more to do with the business decision of the brand. I can say a couple of people have complained but you can’t tell they only happened a couple of hours after you rolled out the new application.

How does your data get used as part of the marketing process?

As with any market research, we deliver insight – we don’t just say that interest is 20 % higher but we try to work out why it is higher, suggesting how they can apply that in their business.

The next step is to say from a social media point of view is what am I trying to achieve – if as an artist or brand you’re trying to engage people to be advocates you need to say: ‘am I going to do an engagement programme, what do I want people to talk about?’ That’s a lot more complicated, it has to be a lot more integrated, work has to integrated with the rest of your campaign, the messages have to be the same but more subtle because you don’t own the space, the consumer owns the space.

The key thing we’ve been focusing on is the issue of advocacy. Am I explicitly saying you should buy it, am I’m telling you what to do or is it an implicit way like “the Citroen Xara is the best family car”- understanding those is crucial to showing correlations of sales, that requires people who are experts at analysing, exploring and understanding buzz – it’s not enough to monitor buzz, you have to understand it…it’s only from understanding that you get actionable insights.

Tags: alan ault, marketing effectiveness, measurement, metrics, music measurement, wavemetrix

This entry was posted on Tuesday, January 12th, 2010 at 2:56 pm and is filed under Interviews. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.