Marketing is now a numbers-driven function. If you're still running marketing campaigns by gut alone, you're behind the times and the ROI probably isn't where you want it to be.

In the last couple of decades, marketing analytics have exploded. It started with web analytics and has spread out to every other area of marketing, slowly but surely.

So instead of embarking on marketing campaigns that have long planning, placement, and evaluation cycles, marketers are now testing and monitoring concepts in near real time so they can replace a badly-performing campaign with one that gets better results.

Still, marketers are struggling to get things right as evidenced by what we see online, in direct mail, in stores, and elsewhere. Here's why.

Single Channel Analytics Are Myopic

There are tools for measuring just about everything related to marketing: tools that monitor certain types of ads, tools that measure social media engagement, tools that monitor customers' behavior on a website, and more.

The great thing about monitoring activity in a particular channel is that one can see what's working and what isn't. Single-channel tools allow companies to optimize what their doing in a particular channel, but they don't provide the entire picture.

The entire picture involves multiple channels.

Multi-Channel Analytics Stop Short

More tools have become available to measure what's happening in multiple channels. Using multi-channel analytics, marketers can see which channels are performing well and which aren't. Marketers use this information to make budgeting decisions.

Sometimes the results are counterintuitive, however. For example, I was recently talking with Chris Madden, the CEO of digital agency Matchnode who told me that one of its clients, New Balance, was driving customers to stores rather that its website because store sales are more profitable than e-commerce sales.

Given the overhead of a physical store versus a website, the result is counterintuitive. Yet, that's what the numbers say.

Importantly, marketers are using multi-channel marketing analytics to better understand the customer journey since most of us do research online, buy online, and buy in stores. The question is, what combination of things resulted sale of a particular item?

Was it the ad we saw on Facebook, a direct mail campaign, an email we received about a promotion or a combination of the three? It turns out that attributing a percentage of a sale to each of those things is still not an exact science.

It's especially hard because multi-channel analytics misses an important point: it's not about the channels, it's about the person.

Omni-Channel Analytics Are Nascent

Some people use "omni-channel" and "multi-channel" marketing terms interchangeably, but there's an important distinction: Omni-channel marketing is about a person's experience across channels and multi-channel marketing is about the channels themselves.

Omni-channel analytics is where marketing is headed, but it's difficult to do when the metrics in one channel are not accurate. For example, if I visit a website using my laptop and then my iPhone, I am counted as two different visitors as opposed to Lisa who is simply using two devices.

The goal of omni-channel marketing is to understand the customer journey so it becomes more obvious what influences which type of people, when and for what reason. The "for what reason" part is called "intent" by marketers.

Intent can be inferred by specific types of behavior, such as searching for "luxury cars" or searching for a Jaguar XK.

In the former case, one is likely researching luxury vehicles. In the latter case, one is likely looking to buy a Jaguar XK, assuming her income and buying behavior align with that (which is information available for sale).

Similarly, analytics may show that the person who checked her Michael's app for this week's coupon is more likely to buy something who hasn't checked her Michael's app or doesn't have Michael's app installed on her phone.

Ideally, we'd like to track the entire journey of individuals to understand what combination of messages in which channels and in what order will make it more likely that a person buys something or takes some other action we want her to take, such as signing up for a loyalty program so we can get more insight into her buying behavior.

However, personalization brings up privacy issues which further complicates omni-channel nirvana.

More Right Than Wrong

Marketing analytics are helping businesses market themselves more effectively, but not entirely accurately. It's the "not entirely accurate part" that consumers seem to remember such as the item you just purchased which continues to show up in search ads and your Facebook News Feed for weeks afterward.

Unless an offer is unusually appealing, effective marketing is invisible. We buy things because we want them or need them and don't think about all the factors that shape our decisions as consumers.

If you're not happy about the ROI you're getting from your marketing campaigns, seek out a good agency or conference that can help you understand how you can improve what you're doing in one channel, in multiple channels, and across channels. You'll likely change your thinking for the better and improve marketing ROI.