Digital Marketing Guide: Analytics

How to Make Sense of All the Data You're Inundated With

By Published on .

Most Popular

Analytics
Help! I'm floundering in a blizzard of data; how do I separate signal from noise?
The first thing to do is pin down your campaign objectives because that will decide your relevant analytics. The second is to look at the data you already have that adds value to the campaign. Most marketers have some of their own data and the second step is to find a partner that can help them integrate that with the publishers. This represents an evolution from the historical dependence on publisher data. Marketers know what segments they want to reach -- if we can support using that data to help target, we are much more likely to meet their objectives. The third thing, after they've told us what they know, is ask, "What can you tell me? Surprise me with some insight."

What's the biggest misconception about data that marketers have today?
Marketers still believe you can't establish ROI online. For direct response or search advertisers, that is not true. But a lot of the large brand advertisers are still very skeptical and in 2010, that's a misconception. Some ask, "Can I get a better return online than my campaigns on radio and TV?" We tell them a couple of things. We (and others) provide solutions that can trace offline behavior and connect it to online advertising.

Is the click still an important metric? If not, what should I be focused on?
You care about the click but you need to demand more in attribution modeling. You need better insight into what we call the assist -- all the many things go into that click. You need to care about the click-stream itself -- you want to know what happened during the nine clicks before the purchase. Did you see a brand ad on any of those pages? We know it makes a difference but sometimes publishers don't share it. The industry should be aligned on consistent attribution modeling. If you give marketers the data, they they can set the weights and how they do the attribution themselves.

Last year, the talk was all about engagement. Does that mean there's a new metric I should be thinking about?
I think the classic metric is time spent -- and time spent is still the best we have today. But all time spent is not created equal. What you want to be thinking about is the quality of that time spent. As an advertiser, especially as a brand advertiser, you want to know how much time a user spent in a quality environment. I don't care if they spend 100 minutes at publisher.com or in a social network, but what about publisher.com/autos? What's missing is a way to clarify time spent in a more-actionable way.

If I'm a display advertiser, is there data I should be looking at in search?
The cause-and-effect relationship between search and display advertising continues to be a hot topic. If you're a brand advertiser and want to target behavioral categories like auto intenders or high household income, you want to ask search partners how those same segments perform on keywords and integrate that with a display campaign. You want to ask them to prove what lift you're getting from buying search and display together. More often than not, search and display is the same conversation. In the past year and a half it has become much more integrated.

Shouldn't display operate a little more like search in terms of using data to continually optimize campaigns?
At Yahoo, we have a slogan: optimize the inner loop. We are continually striving to answer questions, test hypotheses, and make decisions quickly and in a tight iterative loop. The industry needs to focus on reducing the cycle time for generating insights -- the faster we can observe, decide and act, the more effective we will be in our product development and advertising campaigns, and the more value we will create for the entire marketplace. If we do this, effective use of data can be a competitive differentiator for advertisers, networks, agencies, and publishers.

Where can I unearth value in data that my rivals don't yet know about?
It's not about how much data you have, it's what you do with it. If you can just take what you have and clean it up and join it with other sources in a privacy-safe way, that's when you're going to get a leg up.

ABOUT THE AUTHOR
Scott Burke is VP-engineering at Yahoo.
In this article: