Hello everyone, my name is Mat, I’m a paid media director and I work at RocketMill. I’m going to take a moment to talk to you about how we’re in an era of people based marketing. Since inception, advertisers have focused on the media rather than the person. Let me explain to you what I mean by this. Historically advertisers purchase media as a proxy to reach all of their audience and in practice, if they bought an ad in the break of Coronation Street, they did so buying all of that audience in the knowledge that only some of that audience were relevant to their target market. This meant it was vast amounts of wastage. Thankfully things have changed. We are in an era of people based marketing. Using data and technology we can target people and individuals, bidding on them in real time across a wide range of media. So, over the course of the next few minutes I’m going to share with you how I’ve taken a data-driven approach in this era on some of our clients.
In people based marketing, it is vital that we use data to truly understand that audience. So many of us have these pre-consumed ideas of who that audience is, often unconsciously biased. We all need to use data to qualify these assumptions. The demographic report in GA is a great place to start and on the face of it might look simple, but using secondary dimensions, you can really bring this report to life. Not only can you merge your age report with your gender report, but you can also actually start to bring in your enhanced eCommerce data.
So, rather than just seeing the number of transactions revenue and conversion rate a demographic has, you can actually start to see the checkouts that people have used in order to make their conversion. So, maybe a certain demographic prefer PayPal over Apple Pay and you should look accordingly at that in your ads later. Maybe a certain demographic uses your offer codes, your voucher codes more often than another demographic. And if there’s a demographic that doesn’t engage with your offer and voucher codes, then maybe you should look at changing your messaging for that demographic only.
What I really like to pull into this report is the product level information. So, rather than just seeing that one of my demographics has a very high level of transactions, I’d like to know what products they’re actually purchasing. And this can be very, very powerful when taken a data driven approach to our media.
How we can use this approach
I used this exact approach when I started working on the Bauer media account. When me and Lotte worked on this account, we set upon the process of restructuring their whole AdWords account. But in order to do this, we looked at the demographic report first to really understand that data.
What we found out is, despite empire magazine being predominately male focused and the advertising being predominantly male targeted, it was actually females between the age of 35 and 44 that had a much higher conversion rate. Not only that, but by pulling in that enhanced eCommerce data, we were able to see that these women were actually buying gift subscriptions. So, they were buying Empire magazine for someone else. Maybe their husband or maybe their son. And we knew if we could tap into this information and utilise this information, we should be able to increase the number of subs and not just hit our target, but beat our target.
To do this, we use the demographic for search ads in AdWords and it allowed us to take that information from Analytics and actually apply it to our AdWords account. It meant for that female audience we were able to increase our budget and increase our bids, taking a larger percentage of the market share and hopefully increasing the number of subs. But alongside that, it meant when a woman between the age of 35 and 44 searched for something on Google such as Empire magazine or magazine subscription, we were able to show them a tailored ad talking about gift subscriptions.
As a result of taking this data driven approach, we saw an uplift of 42% in the number of subscriptions and a decrease of 36% in the CPA. Now, these were just two elements in a building block of a cathedral of a campaign. And if you’d like to know more, I’d urge you to look at Lotte’s Bauer case study. But the demographic report is just the tip of the iceberg. There’s so many great reports in GA that we can utilise to access the data. And some of those reports are the affinity categories and the in market segments.
Just to be clear, an affinity category is a group of people that share the same interests. So, maybe going travelling or playing football and an in market segment are a group of people that are in market to purchase something. So maybe tickets to go on holiday or maybe boots to play football in. But what is great about this report is it’s in the same style as the demographics report. Meaning you can use those secondary dimensions to pull in information and bring this report to life. For example, you can actually bring in the demographics in this report, getting a clearer picture of who your audience is. Not only that, you can put in that enhanced eCommerce data, so, you don’t just have to look at transaction conversion rates, age and revenue. You can actually see what they’re buying or if their using voucher codes or what checkout they’re using.
Using the source report effectively
When I use this report, what I like to do is put in the source. So, that’s where the user was before they came to our site. And the reason why I like to do that is that, across the social display on YouTube, you can actually buy these segments. You can buy these audiences, meaning if you have an audience that resonates very strongly with a certain platform, be it Facebook and converts well, you can go straight into Facebook and buy that audience.
So, when working on one of our clients recently and putting together a media plan for display and prospecting activity, it was vital that I really understood the audience. And I looked at these reports, however, before I did I already had a pre-consumed idea of who I thought that audience was. I was picturing a 50, 60 year old man smoking jacket on, study, cigar in hand, legs up on the desk reading this publication.
To an extent I was right. It was that guy, however, what I found out is they would spend the majority of their time on dating websites. Now, placing a display ad on a dating site is very left of centre and historically I’d have never considered it, however, based on the data and as long as it is safe to do so, a campaign like this could be groundbreaking for this brand and also give us the edge over our competitors. And this brings me onto segmentation.
When looking at demographic reports, when looking at the affinity categories and the in market segments, some of the metrics I like to look at are the conversion rates, the number of sales, the leads generated, the CPA, but most importantly the value that audience has been to us in the past.
Google is brilliant for this. You can take those segments and you convert them into real time audiences, scoring them on how relevant and how valuable they are to you as a business. However, you’re probably not going to want to use these audiences for remarketing, as these are people that have already purchased and bought your product. When you get these audiences and you put them in AdWords, Google will instantly create a segment of new users based on similar behaviours and similar patterns to that audience of relevant and valuable people that you have uploaded giving you a new pool of people that you can target in your prospecting campaigns.
It doesn’t mean you have to stick within the Google platform. It’s just Google make it easy for us. Most of it is automated. It’s just a case of applying these audiences to our campaigns. However you can take that data out of GA, create list, buckets of those demographics that convert highly and use them across the whole digital media landscape.
Digital is no longer a single channel
Now, I’m not just talking about the platforms that you associate with a digital media agency, the search social media display. Digital is no longer a single channel where you have to wait for people to be online for them to see your ad, digital is a way of trading media. I’m talking outdoor, TV, radio, along with your social search and display.
More and more of our clients are coming to us asking us if we can buy a TV for them using a data driven approach. But before I finish today, I want to share with you one last thing. I’ve worked in media a long time and throughout that period I’ve audited so many accounts and one of the things I look at is how people set up their remarketing. And it surprises me that still today people would just use an all visitors list in their remarketing campaigns. I mean that advertising to people that have bounced on their site, have already converted or people who are just not interested in their product, but they’ll just have this one audience and they’ll apply it to their campaign and say it doesn’t work.
Using a propensity model
I’ve talked about some of the reports I look at and some of the metrics that I use when building my audiences. And this is a great start to improving your remarketing campaigns. However, in order to get the best out of your remarketing activity, a propensity model really helps. Now in, short a propensity model uses algorithms to understand the behaviour of every one that comes to your site. So, all those people that have converted and all those people that haven’t converted. Then when someone new comes to the site, it will score them on how likely that person is to convert based on whether they share the same behaviour patterns or have the same traits as people who have converted or have not converted.
However, this is really hard to do. Unless you have a data scientist in your team, it’s not something that you can just activate, however as we move into this era of people based marketing and the data driven world, machine learning has become more and more accessible and Google are playing their part. They have launched a new conversion probability report. And the conversion probability report gives you your own propensity model in your Analytics account. It analyses everyone that comes to your site. The converting users and the non converting users. And then when someone new comes to your site, it scores them based on one to a 100. A 100 being very similar people that have converted in the past and zero being similar people that haven’t converted in the past.
What’s great about this is because it’s in GA, if you’re using the Google Stats to run your activity, you can take these segments, convert them into audiences and push them directly into your AdWords account. Giving you audiences and groups of people that you can target scored from a 100 to zero based on how relevant they are and how likely they are to convert for you to utilise in your remarketing campaigns, preventing you from wasting your money on people that aren’t likely to convert. And you’d be able to utilise that money on people that are likely to convert.
So, for any of you that have used remarketing in the past and found that it didn’t work or for any of you that use remarketing now and just aren’t happy with the results, I urge you to give this report a go and see if you can change your remarketing. Thank you very much.