Head of Display James Bourner takes a brief look at the data that we use to buy display media on our client's behalf and through programmatic trading.
Hi. I'm James, Head of Display at Jellyfish. We'll just take a very brief look at some of the additional bits of data that we use to buy display media on your behalf and through programmatic trading.
Now, hopefully, you're familiar with the concept of RTB, which we've covered in a different video. But this is a look at some of the data that powers the decisions, powers the algorithms, and helps us buy exactly the right media for your particular campaign.
The easiest way to look at this is to follow an ad call, follow the journey of an ad call as an impression becomes available in a page.
So, generally, pretty much always, what happens is a user will visit a page. That page has advertising on it, which is made biddable through ad exchanges. The webpage will call the ad exchange. The ad exchange will package up or announce that impression to people like us that sit there and listen to these impressions coming through exchanges.
The ad exchange appends some information to this impression, such as time of day, day of week, the site, the browser, information about the user's computer, nothing personally identifiable. That's illegal, which is good but also sad. It allows us to start to make a decision about the value of that ad impression, whether we want to buy it for your advertising campaign, and as well, how much we want to pay for it if this is a true RTB part of the campaign.
So to further enrich or to help us make better decisions, we won't rely just on that little bit of information that the ad exchange passes us, but we'll go and check with other third parties, but most importantly a lot of first-party data from yourselves as well. This is one of the most overlooked things in display advertising - the power of first-party data.
So this impression goes into the ad exchange, and the ad exchange announces it. Our DSP is listening to this and making decisions. While it's making that decision, it'll quickly go away and check. Imagine we've set up a campaign, and there's an awareness part of it where you've asked us to purely buy demographics, let's say females between the ages of 34 and 54. There's also a retargeting element to this campaign as well, where we use your site data to be more aggressive with that user because they may have fallen out of your purchase cycle or not.
Additionally, we have to look at brand safety, and every impression that we ever buy is subject to this as well, because it's very important because the Internet, of course, being a wonderful place, is also a very colourful place and most of that colour is not suited for anybody's brand. So we check that every single time.
So what happens is thus. This impression becomes available. Our DSP very quickly goes and checks against a third-party data provider. It could be somebody, like Experian, eXelate, Core Credit, VisualDNA, who can provide us real data from offline sources, which will give us the age of the person sitting behind the computer, the sex, their income, whether they own a house or not, not personally identifiable, of course, but whatever that cookie matches.
Then there's other providers that can give us things like intent behaviour. So if somebody's been looking at lots of car websites and car reviews of medium-size hatchback, these companies might decide or infer they're in the market for a car and, because of that model of car, that maybe they've got a couple of kids or something like that. That's the third parties that we'll draw on. So all the time, we're cross-referencing those, and that's affecting the price of the bid that we want to put in.
Now, notwithstanding that, it's the first-party data and, again, the most important part. So this might be a returning customer to your site. This might be a customer that's falling out of the purchase funnel. Therefore, a remarketing budget would pick it up, and we have messaging accordingly.
As well as that, because of the integrated stack - and this is what we passionately believe in here at Jellyfish - is they may have been a customer that searched for your brand terms and been to your site or not been to your site, but, because they searched for your brand terms, we'll start to know that there's no point in showing that person an advert from your awareness budget. We'll use prospecting budget or retargeting budget to reach them.
So you start to sew together, seamlessly, marketing plans, which is something we've been talking about for 10 years. But I don't know, honestly, I think it's been a little bit flaky. So, notwithstanding that, we can also go into CRM data if we need to. So we can start to re-message people based on whether they're high net-worth customers, or if they've bought a pair of shoes, why not this handbag, or even notifications about a sale or upcoming event on your store.
This is why, when you start to sew all these things together, your marketing money works harder, which is why we totally believe in operating like this. What's more impressive, all of those queries, all of those providers, all of those bits of data here, one hundred milliseconds we get to make that bid, tenth of a second. Otherwise we lose it, and that's why it's good!