Cause and Effect

One of the inherent problems with marketing things on the internet is that some bits of the process are very very easy to measure. That's not a problem for people who sell things online, as they can easily understand how many of them were sold and at what cost. The problems start to arise when the things being marketed are either sold elsewhere, and even more so if they are not intended to be bought immediately: if for instance the goal of the marketing is to change people's perception of a brand. Hence banners are still regularly measured on click through rates, even if the product they are advertising isn't for sale on the page being clicked through to. Clickthrough rate is clearly to wrong thing to measure in this instance but it is EASY to measure, so often taken as a proxy for less tangible results.

As brands become more interested in the social web, there is a real danger of making the same mistake again. As for many brands the social web currently means social networks (and for a large proportion of them social networks largely mean Facebook) the easy measurement tends to be numbers of Likes (/fans/followers/etc). Social networks are where people are spending more time, so brands want to be there. To justify being there, they need to measure something. Numbers of fans are easy to measure as brands and their agencies can plan, track and optimise their activity over time, and because Facebook has a straightforward advertising platform they don't have to think differently to how they always have done.

The problem with using these numbers as an objective is that no-one knows what value to attribute to them. You and I know that attributing a single value to them is meaningless, as every brand will be investing in community for different reasons and starting from different places, but that hasn't stopped a range of 'one number' valuations being put on Facebook fans. I'm going to ignore the Nielsen research published a few months ago, as I'm sure you have too if you have read more than a few lines into the introduction, but it's worth reading The Ad Contrarian's analysis of it. Syncapse's 'Value of a Fan' study is more interesting because unlike Nielsen it is genuinely empirical research - it starts from the premise of trying to things out find out, rather than trying to prove them. It also accepts that all brand communities are fundamentally different so any 'one number' can only be an average.

So in ambition and objectives it seems sound - and is worth a read (download from here). But then I get very confused by the methodology. This is a summary of the findings:Or alternatively have a look at the higher res image here. The study looks at money spent on a brand by fans and non-fans, and find that fans spend more money than non-fans.

Let's stop and think about this for a minute. A major research organisation actually thinks that it is worthy of our time and attention to know that if you like a brand, you will spend more money on it than someone who doesn't like it as much. Or put another way, brand marketing for the last 40 years works, but it's only now that we've got a survey of 4,000 Facebook users that it can finally be proved.

OF COURSE people who like a brand spend more money on it. That is A. not news and B. nothing to do with Facebook. Attributing that as value to Facebook fans is a basic misunderstanding of cause and effect. People liking brands has any number of CAUSES, of which what they do on Facebook is a tiny part. One EFFECT of their liking the brand is that they choose to sign up to receive more information from the brand, or to display their affection to their friends (this in turn may cause more people to interact with said brand, but that is not the subject of this study). Quite often people liking a brand is a result of great advertising - the best advertising makes people fans for no other reason than because the ads are so good, and one of the places that this is reflected is in social networks. Old Spice and Compare the Market saw their Facebook fans numbers increase as dramatically as their market share in response to great advertising. Both increases were therefore EFFECTS of great advertising. But to take one effect and infer a causual relationship with the other is plain wrong: so if instead of more bits of culture, Old Spice bought a load of 'become a fan' ads, this would not have the same impact on their market share. The implication of causal relationship made by this study is no different to saying that for instance:

When it snows, there are more road traffic accidents
When it snows, my toes get cold
Therefore by warming my toes up, we can reduce deaths on the road.

This confusion of cause and effect starts to lead marketers into dangerous territory - if one number can be applied to fans without questioning whether it means anything, then there is genuine justification in buying them - either by soft methods like 'Become a Fan' ads, or by harder ones like buying them on Ebay (no seriously - you can buy Twitter followers on Ebay. Fuck knows why. Check this pic courtesy of @sparkey).

Now there are lots of ways that Facebook is a great set of tools for marketers, but by buying into the cause & effect confusion we are in danger of missing the most important ones.

People who trust advertising

Comscore, Tacoda and Starcom published a research project last year that questioned the clickthrough rate model as an effective metric for online display advertising by showing that 50% of clicks were made by 6% of the online population. These so-called 'Natural born clickers' were not reflective of the people that the advertisers responsible for online display campaigns were generally interested in communicating with.

I was reminded of this when reading Ted Shelton's post on Replacing Anti-Social Marketing, which points out that only 13% of people trust advertising (according to the Edelman Trust Barometer). Ted raises a really important question from this - are these the people that your brand wants to be seen with? Given the range of other information available today, would you really base a business strategy on people credulous enough to still trust advertising?

Online Advertising 101 - This Ad is Annoying

Advertising with my own money....
With all the possibilities that have opened up over the last few years for brands to join in conversations, I sometimes forget that the economics underpinning the biggest social networks are still all about eyeballs, scale and broadcasting. So, selling display ads! I saw a presentation earlier this week talking about the success or otherwise of running traditional advertising on Facebook: now most people would say that this isn't the best way of using Facebook as an advertiser, but while there are millions of people using it there will also be lots of DR campaigns chasing their eyeballs and wallets. Interrupting people with commercial messages while they are chatting to their mates is no different from selling cheap DVDs table to table in the pub, but there is clearly money to be made in that as well, because people do it. And it is the same on Facebook: if the ads are cheap enough then it doesn't matter how few people notice them.

And they are really cheap - actually really really cheap: starting at around $0.10 CPM (as all Facebook ads are priced in US$). Now don't have much experience of direct response text ads on Facebook, so when I heard this, I thought it might be worth while having a go.... put something provocative together, stick $15 behind it, and that would get me about 150,000 page impressions.

So what do I have to advertise? Well, it might also be useful to see what the people who noticed the ads thought of them, so I came up with this:

This ad is crap.
Does crap advertising on your profile page annoy you? I work for an advertising agency that puts them here. Tell me where we're going wrong.

So a lucrative sideline in copywriting clearly awaits.... Not surprisingly this wasn't accepted, because those polite souls who run Facebook don't allow offensive language in ads (this surprised me at the time considering what content is on Facebook, but what doing this has taught me is that each user of the site will set the content at a level they feel happy with by choosing who their friends are. Advertisers want to reach everyone, so have to be more careful. More on this later). So I changed it to this:

This ad is annoying.
Does annoying advertising on your profile page annoy you? I work for an advertising agency that puts them here. Tell me where we're going wrong.

And I set up a WordPress blog called This Ad Is Crap as a landing page giving a bit more information. The rationale for hiding my identity, and claiming in the personal information that this could get me into trouble if I was found out was to try and create a bit more of an undercover feel - that anyone reaching the page might feel more like they were part of a secret project. (rather than being part of a cynical piece of research - not sure if my rationale is correct or ethical, but that was the reason).

After having to remove all references to Facebook from the landing page ('breach of copyright' apparently), my ad was accepted. I targeted all adults 18-35 in the UK, and bid $0.10 CPM with a daily cap of $5. In retrospect I should have spotted that there was a CPC option as well, but this was a bit of a rushed job...

And sat back to see if anyone noticed it. My view is that this is a slightly different ad from what you normally see there, so I was expecting a 'good' clickthrough rate. 'Good' in these sort of environments means 0.01% (or an Ignorethru rate of 99.99%). So a click target of 15 ($1 per click). However my inept attempt at a landing page, and the fact that I was asking people to offer their thoughts in the comments, would probably mean very few actually commented. So these are the results:
29 clicks, 167,582 impressions, at a CPC of $0.51. And one comment. Considering lack of image in copy, dodgy landing page, all the online adverting 101 stuff that I would tell anyone doing it for the first time and didn't get round to myself, not a bad job. This shows why so many direct response advertisers are there, as it is so very cheap that it doen't matter if hardly anyone actually notices the ad. If you actually have something interesting to say or sell (rather then just being a smartarse like me) then this is a great system, particularly if the targeting is accurate: I didn't make too much use of targeting here, but that might well be a test to play around with next year if I ever have $15 going spare in 2009.

It is also interesting to think about when talking to people who think their 'web 2.0' project on Facebook means that they are participating in social media. I now have a case study paid out of my own pocket that says that just catching someone's attention while they are with their friends with an ad that is a bit different doesn't mean that they will the spend any time joining in with what the destination site wants them to do.

And although one comment is very far from representative, it is certainly an interesting perspective, and one that has made me reconsider my cynical reaction to my initial ad not being accepted.

Anyway, back to spending other people's money armed with a few new opinions.....

Opportunities not to see….

The rest of the family has been away for the weekend, and it’s made me realise how easy it is to write when the TV isn’t on (ok, it’s made me realise how much easier it is to write when there isn’t a baby trying to eat the laptop as well, but anyway…). If I lived in a BARB household, pretty much everything I write would be whilst I was being counted as an active TV viewer. I believe the technical Touchpoints term is concurrent consumption (ie two media channels at once). So I was in a research meeting today, talking about the relative value of TV impacts , which has historically been higher than other types of media. For example, Millward Brown studies over the years weight magazine impacts at about 60% of the value of TV impacts, because although ads in mags are more likely to be contextually more relevant than TV ads, there is also less likelihood of you actually seeing one if you have just opened up one issue. And because BARB is a more detailed and accurate survey than NRS, etc. This is before you factor in how audio-visual content has proven over the years to build emotional bonds with brands.

So that got me thinking about how we (that’s we the people, not we the media planners) are viewing TV. This is a moot point at the moment, because on the one hand TV has never had better content, never been cheaper in real terms, and in all probability will be watched even more next year when we can’t afford to go out of the house. But on the other I don’t know anyone who watches many TV ads any more. I think this is the underlying factor behind Tess Alps’s increasingly
frantic repeating of all the pluses for TV advertising that I mentioned earlier, to remind the Guardian Tech readership that, yes, some people do watch telly still. Because there is no avoiding the facts: audio-visual ads still build brands more effectively than other traditional advertising formats, broadcast TV is still the quickest route to play those ads, and it is cheaper to do so on broadcast TV than ever before.

….But….

what if more and more people are watching all of the hugely diverse quality content on TV, but not seeing the ads? Obviously we know there are plenty of Sky+ and V+ boxes out there – 17% of the country at the last count (although as the last count was made by TGI, it WAS 6 months ago). Now received wisdom is that having one of these devices means you watch more telly. I know I do. And that while you obviously record some programmes and probably skip the ads in them, the increase in amount watched is more than the total amount timeshifted. This all comes from BARB data (admittedly with a Thinkbox spin) from 2006.

Trouble is, if you test this theory on unsuspecting members of the public, they tend to laugh at you. There’s a fundamental disconnect between UK qual and all US research , which says that we don’t watch TV ads, and UK quant (and this is all BARB stuff, so proper sample sizes), which says it isn’t a problem that we need to worry about at the moment except for a few must watch dramas that are disproportionately heavily timeshifted (but where we can buy cheap ads in the on-demand version anyway). So how come there is such a perception gap?

Well the main reason is VOSDAL. Viewed On Same Day As Live. A piece of BARB terminology that dates back to the VHS era, which was designed to add as many timeshifted viewings (you remember the kind – put chunky tape in recorder, fight for half an hour to programme the thing, keep the tape for months after you’d watched it because you didn’t want to delete something that had been such bloody hard work to get onto the tape in the first place) into the overnight figures that TV buyers managed their campaigns around. All timeshifted viewing is still reported back now as either same day or later

Thing is, it isn’t that difficult to record a programme now. Personally, I like live viewing, because you can use the EPG to skip around. But if the phone rings, the baby cries, I fancy a cup of tea, whatever, I’ll hit pause. Because I can. Actually, it works even better if it’s on a commercial channel, because then there will be 9 minutes of ads every hour that will let me catch up with live again, to get my EPG back. And I'm not the only one. Everyone who has Sky+ does this. That's what it is designed for, and it is what the Sky+ ads are all about



In fact ideally, we need VOSSAL – Viewed In Same Second As Live – to have confidence in our ad campaigns. So I had a look at BARB’s FAQs and methodologies to find out what counts as VOSDAL today. Please follow the links if you can be bothered, and if you find any evidence that it has changed from the VHS days let me know, because I couldn’t. In fact, when I rang BARB, their helpdesk couldn’t tell me either. All they could do was assure me that the overnights were accurate. Now BARB is statistically extremely accurate, so I don’t doubt them. But I don’t think that they actually measure if DVR owners are watching ads. And I suspect that this might be why I get so suspicious that Thinkbox have got something to hide when they tell us that TV has never been better. It hasn't of course: we have the best telly in the UK now that any country has ever had, and it will get even better when Kangaroo launches. But the audience is up to something, and I'm not sure if BARB are able to monitor it.