How digital advertisers can make the most of their data
With a little effort, advertisers can get much more out of data than they put into it, says AudienceScience's Mark Connolly.
Big data was one of the big topics of conversation last year and I'm sure it will be again this year. What I wonder, though, is how many people will move from talking about it to actually doing something with it?
On the surface, it seems as though everyone is creating magic from big data, but the truth is that most people in the industry are still trying to figure it all out. Many lack even a basic understanding of this data, much less the terms used to describe it.
The important point to recognise is that how you use your data is just as important as having lots of it. You need data to be scaleable and so must have a clear plan about how you are going to collect it and segment it, etc. Then you need to have the right tools and staff in place to make it actionable, with enough data volume to feed your marketing activity.
The truth about data is that it's only valuable if you can obtain actionable intelligence from it."
However, to really help make sense of big data, let's first define the three kinds of data that are generally discussed in online marketing.
First-party data: This is a company's own proprietary data assets, collected from online interactions, CRM, transactions, social media, customer services, etc. This data can inform marketing activities and facilitate communication with these customers.
Second-party data: Used like first-party data, second-party data comes from a partnership with an external party, often a publisher. That external party is the direct source of the information and they agree to share data with strategic trusted partners such as advertisers. Advertisers can pay for this data, but it is often shared on a quid pro quo basis.
Third-party data: Data obtained under license from a third, external party, usually an organisation that can assist with building segments and finding audiences online. Third-party data may come from behavioural data companies, shopper data companies, offline data warehouses or financial data brokers and is commonly used to prospect new consumers.
The next step is to dismiss a major misconception about advertising data: that amassing tonnes of it will make a brand smarter than the competition. The truth about data is that it's only valuable if you can obtain actionable intelligence from it. It's called "big data", but in reality lots of small insights are what help target campaigns with more precision.
What advertisers need to do first is make the most out of what they already have - the first- and second-party data. FTSE 100 brands often have online profiles created by their customers, or offline data from loyalty cards and frequent-buyer programs. That's extremely helpful for reaching customers who have brand awareness and are very likely to purchase the same product again.
This kind of targeting can be strengthened with the addition of second-party data from a trusted publisher. Sites like Yahoo, MSN, and major newspapers know who visits their sites and what kinds of content they consume. Slicing these data sources in different ways provides a great deal of insight while remaining cost effective for advertisers.
Advertisers and their agencies will need to keep adjusting the data to find the right mix."
But first- and second-party data, while affordable and effective, carry limitations. For instance, the advertiser is only targeting consumers that already visit their site or buy their products, meaning these consumers are likely to buy the same brand again. Not even the biggest advertisers in the world have 100 percent coverage of every online consumer, and they certainly don't have the same detailed profiles for each and every customer.
Let's take an example: a brand may know who buys their product based on an anonymous online profile the consumer created, but it doesn't know much about the consumer's interests, behaviours, household size, or estimated income. Third parties can infer some of that information, in a data-protection-friendly way without any personally identifiable information.
Third-party data expands the reach and is often used in combination with other sources to introduce campaigns to new customers: Lookalike modelling is based on first-party data, but identifies new prospects. Third party data partnerships should always benefit the overall data pool, by adding scale and depth of understanding to the available first-party data.
Once armed with data, deciding which sources to use and how to leverage them comes down to each individual campaign and its desired outcome. Different brands will have different needs depending on their target audience and specifics around how people purchase their products. For instance, an automotive brand needs to reach consumers on a long purchase cycle. That requires refined targeting, knowing not just that a consumer is interested, but when they are interested.
The final element in optimising the use of data is to analyse campaign results, look at the money allocated to data and ensure there are clear benefits. Transparency around data costs is important here, as it allows brands to see if their investment is worth the return.
Advertisers and their agencies will need to keep adjusting the data to find the right mix, but the more they study, the better they'll perform. With a little effort, advertisers can get much more out of data than they put into it - the focus shouldn't be on acquiring data, but getting more out of it.
Mark Connolly is chief revenue office and vice president of international at AudienceScience