I know what you’re thinking: This seems a bit premature. A 2016 article at the start of 2015? It’s not a typo, though, and it’s deliberate.
There are copious 2015 planning articles circulating at the moment — but if you’ve not started planning for 2015 yet, then you are probably too late to do anything big. Sorry.
And “big” is what I’m talking about here. What large-scale tasks do you need to complete in order to take advantage of everything that’s coming your way?
For major payoffs in 2016, you need to put extra effort into these areas over the next year:
Let’s look at each in turn.
Clean data sounds simple. It sounds like a hygiene factor that should be assumed. In fact, if you ask the teams that look after your database, I’m sure they’ll tell you the data is already fine. It probably is, for some purposes. But we’re talking about your digital marketing, where requirements can be different.
Think about how many data-driven tools and platforms have been introduced in the last year: new Google Shopping requirements, AdWords ad customizers, dynamic content feeds for display ads and emails, and the list goes on.
Your data feeds control your targeting, your ad messaging, and even your bidding. You can manage large portions of your campaigns just by manipulating the content of feeds, and huge amounts of that should be automated straight from your website and database.
If you have 100,000 SKUs in your database, it’ll take a lot of work to make sure they all have every field entered correctly. No typos, “pink” instead of “fuchsia,” standardized sizings, for example. These all make a big difference to your discoverability when these terms are being used by platforms to create keyword lists and other targets.
On top of your own product data, your existing marketing data can always be cleaner. Have you categorized your remarketing lists based on their traffic source? Can you do search to display retargeting by search intent? Can you do it by social source?
Making sure that your tagging is amazing and that your remarketing lists take full advantage of that tagging is a great all-year-resolution. Trust me, this will take a while.
By cleaning your product data you’ll be ready to maximize the use of the existing data-driven platforms, and by creating strong data structures and hierarchies you can make the most of your marketing data.
Once you’ve got all that data, you need a way to share it.
Feeds are the obvious first step. Make sure that your ecommerce platform can appropriately output whatever feeds you need. They’ll tell you they can, so put it to the test. The requirements for any digital marketing platform’s feed intake can usually be found in their help files. Check that you can produce a feed that can:
As well as your feeds, you need to be able to share your website data and your marketing data. You have several options for this, ranging from DIY through to enterprise.
The most basic solution is Google Analytics. Chances are you already have this tracking your site, and at no fee you can track everything coming into your site. The sharing mechanism here is in Google Analytics’ remarketing lists. Set these to put people into lists based on the criteria you already created, and you can use it in Google products.
If you want to go a bit further and use them in non-Google products, you need to be able to use their own remarketing lists (or custom audiences, if you prefer).
Google Tag Manager can help you to set these up based on similar criteria to Google Analytics. Use query strings in detail, and use macros to trigger the remarketing tag to fire when you need it to, and get every data hierarchy you need into every platform you need it in.
The top level setup is to use a Data Management Platform (DMP). These dedicated services work wonders for data sharing.
Think of one as a set of relational databases, that link to almost any service or API you can imagine. If your CRM contains Name, Email, ID and Address; and your website browsing data contains page views and ID, then the DMP is the place that links those two datasets together, in this case by ID.
Import data from almost any system, link it via any and all pieces of information they have in common, create valuable segments of users, and turn those segments into targetable remarketing lists. Then do lookalike modeling to find new users who match. Awesome!
We all use first-party data a lot, and we use third-party data where we can, but we’re focused on aggregated, anonymous information. We’ve become so obsessed with the statistical significance of the data we’re buying that we forget about its quality.
The best way to be sure of the quality of your data is to know where it comes from. So think about other businesses you work closely with, and whether you can negotiate data-sharing deals directly with them.
I love this word — partly because it’s so ridiculous and partly because it helps search-centric people think about the convergence of search, social and display from a broader point of view.
All three platforms allow programmatic buying, and all three allow dynamic content in their ads. They each have some target options unique to themselves (keywords, placements and social profile) but share a lot as well.
Consider that, by 2016, you’ll be treating them all similarly and managing them on the same (or similar) platforms. Make sure that your teams are set up accordingly — no silos, shared skills/training where possible, and every optimization session should consider all of them.
Attribution is still as difficult as it was when we started talking about it. It’s rarely done well — in most cases it can’t be done well — and there isn’t much on the horizon promising to make it better.
Cookies are a limiting factor that can’t handle offline and can’t work across devices. User IDs of the kind Facebook have are a limiting factor that are non-standard, walled gardens and still can’t work offline. As soon as any user’s path to conversion deviates from your stereotypical funnel (and most do) the challenges to modeling that behavior become very tricky indeed.
So what can we do? Let’s change the conversation and change expectations. Instead of attribution modeling, let’s talk about closing the loop. We know that we need to get the data about the value of conversions back into the media channels that drove those conversions, and we know that we need to account for the fact not every touch point was worth the same.
If we stop thinking about trying to solve the problem, and start thinking about improving our current position, we break away from trying to say, “This is what’s really happening,” because we can’t get that right. Instead, we want to say, “This is what I know now that I didn’t before,” and we’ve closed the loop a little bit better.
Obviously, this list is entirely made up of things that can be done now. For the advertisers at the absolute forefront, this was a 2014 list. But almost nobody does all of these well, and they all take a long time to set up.
It’s too late to get these all implemented in time to be at the forefront in 2015, but you have twelve months and a strategy so that you can be there by 2016.