The Digital Marketing Stack: Analytics
Ian Lurie Jun 1 2015
In part 2, I talked about infrastructure and its place in the marketing stack. In this post, I talk about analytics and its role in internet marketing.
Analytics measures infrastructure, audience response to content and channel performance.
If you don‘t measure, you can‘t observe. If you don‘t observe, you can‘t adjust. That leaves you jabbering away at your audience with nary a glance at their facial expressions.
If you are the only company that sells the only solution/product to the audience, maybe you don‘t care. And, if you happen to be that mythical company, let me know so I can write a book about you. Everyone else, pay attention.
Analytics doesn‘t have to be hard. Here‘s what you need to measure, and how:
What to know
You need a context in which to interpret your data.
You can‘t function without goals — real business goals —. There are only a few types of organizational goals:
- More revenue
- Lower cost
- Greater response
Figure out the one you‘re after, then assign a number.
If no one will give you goals, pick some, assign outrageous values and share them with your co-workers/boss/client. You‘ll be amazed how fast you get useful feedback.
Pick your tools based on the goals you‘re trying to measure. Never lose sight of the goals! Otherwise you‘ll waste time and money. Typically, though, you‘ll need:
- A web analytics tool. Google Analytics is my go-to, but there are lots of others. Just pick something. Use it consistently.
- Offsite data. Social shares, comments, reviews, sentiment and any other information about user behavior when they‘re not on your site. If you can‘t get this, I understand, but really, why? At a minimum, use a tool like SharedCount. It‘s free. Buzzsumo is a nice choice, too.
- Excel or greater. No matter how good the toolset, you‘ll need to crunch numbers. Excel is my starting point. Fancier tools include R, any SQL-driven database or pre-built toolsets like Tableau.
What to measure
With tools in hand and goals in mind, you need to measure progress. The only kind of measurement that matters is progress towards the goals. Everything you measure must somehow connect back to your goals.
Any time analysis paralysis sets in, come back to the goals. Which metrics most directly measure progress towards the goal? Use those first. Then use the less-direct stuff — things like click location — to do a more detailed diagnosis and plan next actions.
How many people come to your web site? Look at unique visitors, repeat visitors and new visitors.
Web traffic sources
Where do these people come from? You need to know source and medium: “Google” and “organic search,” for example, or “Facebook” and “social media.”
Web traffic behavior
What do they do when they arrive at your site? Look at the pages they most view, conversion rates for things like sales, e-mail newsletter signups, etc. and any other data you can get your hands on.
Minimum, though: Pages viewed and conversions. If you can‘t measure these you‘re chasing your tail. Assuming you can see it. Set up behavior tracking.
What devices — mobile, desktop, tablet, operating system — do visitors most use? Track trends and changes. Your audience is moving towards mobile, fast, and you need to know the unique behavior of users on each device.
How long does it take for your site to appear in visitor browsers? Google Analytics and other web analytics packages can track time-to-load and rendering time. That gives you a real look at the user experience.
Click location and scroll distance
Know where people click on a page, and how far they scroll down that page. That tells you which parts of the page attract attention, which can help you place calls to action and other content.
I like Crazy Egg for this job.
Cost per conversion
All marketing delivers some kind of ROI. You can‘t measure all of it, but you can measure a lot. Make sure you know cost per conversion.
‘Real‘ business metrics
‘Real‘ metrics are unfiltered data pointing back to your goals. ‘Unfiltered‘ means it hasn‘t passed through web analytics software. It‘s the stuff your accountants, accounts receivable and finance team look at. This kind of data includes:
- Foot traffic to stores
- Call volume
- Sales volume
- Cost per sale
You need this data to:
- Measure otherwise untracked trends
- Check against data from other tools
- Connect offline to online trends
- Put business performance in the context of web performance and vice-versa
Even if you‘re an e-commerce site, you have sales and revenue data that doesn‘t appear in your web analytics reports. Those are ‘real‘ business metrics.
Decide. Measure. Adjust.
There are an infinite number of metrics you can measure. Start with the core business goals and build out.
Then assemble your tools, start tracking and add metrics/tools to get even more insight.
But no matter what, start with the business goals.
Note: Read a bit more about analytics and compare it to other parts of the stack in our Marketing Stack Explainer.
CEO & Founder
Ian Lurie is CEO and founder of Portent. He's recorded training for Lynda.com, writes regularly for the Portent Blog and has been published on AllThingsD, Forbes.com and TechCrunch. Ian speaks at conferences around the world, including SearchLove, MozCon, SIC and ad:Tech. Follow him on Twitter at portentint. He also just published a book about strategy for services businesses: One Trick Ponies Get Shot, available on Kindle. Read More