“The future belongs to those who use data.” That quote stares at me every day on a poster as I leave the tube station on my way to work. The poster is advertising a big data conference, mostly focused on technology. When I see it, I ask myself the question, "how?"
Exactly how does the future belong to those who use data? We know that data is valuable - we know that people, companies and governments are creating and collecting more of it all the time. What we don't understand very well is what the actual business model is as a result and what the risks are of building a data-driven business.
It's clear when we look at a companie like Amazon how it makes money from data: it holds and processes on behalf of clients. But why do those clients pay? How does a non-data business start recognising that data is valuable enough to pay somebody like Amazon?
The value of your audience to yourself
The most obvious starting point is how data about how people interact with you is valuable for your own purpose.
Analytics: analysing your web and/or app traffic can bring you immediate returns. If you pay a lot of money for a search keyword that doesn't perform, this is your route to work out why not and fix it. If you've tried and can't fix it, you can stop running that keyword and save yourself some money. Good analytics forms the basis for everything else. Analysing your marketing campaigns will immediately deliver potential optimisations that can save money or drive more revenue, adding value.
Insight: this is separate from analytics, and concerns how you use the data you've collected. It's quite nebulous, but I define insight as something new and useful that I've learned as a result of poking around my data. Those two qualifiers are important: new and useful. Insight might tell me about consumer trends I can use in my buying team. It might help me get a better understanding of my audience to improve my TV buying. It might show me a major competitive advantage I hadn't been exploiting. Whatever the data tells me, it's not insight unless it's information I didn't know before and leads to genuine value for my business.
Segmentation: the better you understand the people who make up your audience, the better you can service them. A data management platform (DMP) is the easiest way to do this, but what's more important than the technology is the set of rules about how you segment your audience based on their behaviour. If a person reads five car reviews, you can assume they're more likely than the general population to be in the market for a new car. If somebody has visited your site twice and checked out your enterprise-level solution, then that's a signal about what they're interested in. These rules create nodes of information about users and each node adds to the others to build a picture of groups. A person might affiliate themselves 80 per cent with one of those groups and 20 per cent with another.
Lookalike modelling: once you have an understanding of groups that your users align closely with, you know to which groups you should show ads. Lots of media sources have a nice, rich understanding of their audiences, so the better you know which of your audiences are your best customers, the better you know which elements of your media partners' audiences to buy.
The value of your audience to your partners
You may be a data owner as much as you are a data customer. If your website has a lot of content, then a user's behaviour can tell you a lot about their interests.
Serving ads on your site: if you host adverts, then your inventory is worth money to advertisers. Your data is worth far more. You can sell space on your automotive section to a car brand, or you could sell space anywhere on the site, but served only to users who regularly read the automotive section. The latter offers advertisers a wider range of inventory and better targeting of relevant readers. You can charge more per ad, serve fewer ads and sell the rest of the inventory at the new rate to other advertisers. Ads sold in this way might cost up to four times the price of inventory that doesn't come with any audience targeting, if those users are particularly relevant or valuable.
Relevant content on your site: the better you know what your audience likes to consume (articles, blogs, videos, product comparisons, etc), the easier it is for you to create more relevant content (improving the experience, improving user retention, improving SEO rankings, improving shareability, and improving visibility). You can also use that understanding to create partnerships with other content creators to serve their content on your site. This gives them the chance to serve the right content to relevant readers, while you receive additional relevant content to improve your on-site experience.
Selling your data to the marketplace: some advertisers might want to be able to use your understanding of users and connect it to ads served on other sites. You can work with a data broker and set your DMP or tag management to fire their pixels when a user is added to a node. When that criteria is met, the third-party data company can cookie the user which is then available to buy against elsewhere around the web. You can charge a commission based on how recent the data is, how certain you are about the data and how valuable the user appears.
Maximising your data's value to the market
To keep your data as valuable as possible, simply work on the things you'd want to know before you use it to serve ads.
How recent is the data? If a person's behaviour on your site indicates they're in the market for a new smartphone, it's safe to assume they won't stay in market for long. Charge more while you can and let your price tail off as their “in-market” status signals age.
How certain are you about the data? How many page views in a category did a user exhibit? Did they break the minimum criteria by 100 per cent? Do they fit two nodes you believe to be closely related to each other?
Each of these circumstances would give you more confidence about your segmentation. Even more valuable than implicit data (behavioural) is explicit data. Ask people. Run surveys, do quizzes, incentivise feedback. If a person has told you a piece of personal information, that is more certain and therefore valuable (assuming permissions are given). Charge more for any explicit data.
How valuable is the user?
If a person reached your site by searching for “cheapest furniture”, that's a very different signal from a user who searched for “painted wood bespoke kitchen”...Whether it comes from searches, from browsing behaviour, from filtering/sorting behaviour or from the prices of items viewed or added to baskets, there are a lot of data points that tell you whether a user has a large budget or a small one. Anybody with a high value rating should be in a different audience pool, one you can charge more for. These pricing differences should be applied whether you're serving ads on your own site or selling the data segments to advertisers to be used elsewhere.
Your data is valuable, so collect it, analyse it, use it and where necessary, monetise it.