Ryan Jones is a wicked smart SEO who works with SapientNitro and rants on his personal blog (dotCult) about all things SEO, Penguin, and analytics. He’s one of the most “outspoken” marketers we know and Rhea will be moderating a panel with Ryan at the upcoming SMX East conference, October 2nd-4th in Manhattan. We hope to see you there! In the meantime, Ryan shares what you should be measuring and why.

When somebody comes to me asking for data I always ask the same questions: Why? What are you going to do with it? I don’t enjoy giving people a hard time [too much] and I’m not [that] lazy. I just want to make sure that they’re asking the right questions and answering them with the right data.

If you take one lesson away from this rant, let it be this: If you’re not asking questions, you’re not providing as much value as you could be. Chances are you were hired for your expertise, experience, knowledge, and insights – not simply your ability to sort numbers in excel. The person asking you for data has expertise, experience, knowledge, and insights too – but if it’s not in your field then it’s your job to ask these questions and ensure you provide them with the right data to make the right decisions.

A good analyst needs to understand not only what they’re measuring but why they’re measuring it too. Proper analysis demands it.


It’s not enough to just measure what’s easily measureable; we also need a reason for measuring it. If there was a formula for increasing web sales our jobs would be easy, but there isn’t. The closest we come is that dreaded correlation word everybody keeps throwing around. When we apply statistics to our website data we can get an idea of things that correlate highly with sales. We can then focus on increasing or optimizing those “things.” It’s here that lots of companies run into trouble. I’d like to use hypothetical (but based on a true story) example to illustrate just what can happen when people don’t ask why.

Imagine the following scenario: (Note: I’m making this up, but I see this type of thinking all the time.)

A senior executive learns (correctly) that Youtube views correlate highly to sales.The latest correlation data says so.   The executive sends out the order but only includes the ‘what’ without the ‘why.’ “Hey, get us more Youtube views. They’re important,” he bellows, and his words echo throughout the company. Wanting to impress their boss, managers incorporate Youtube views into their bonus metrics and make them a priority.

Fast forward a few weeks and suddenly paid search ads are now pointing to Youtube instead of the company website. The company homepage is driving people to Youtube instead of into the shopping cart. Product features and testimonials have been replaced with embedded Youtube videos. Youtube video descriptions now call out more videos from the company.  Sales are down and mid-level managers are screaming that we obviously need more Youtube views.

It sounds absurd, but I’ve actually seen this exact scenario happen at some very large companies (thankfully, not so much within my company.) Go ahead and substitute Youtube views with  facebook likes, pagerank, mozrank, alexa rank, twitter followers, retweets, reviews, ratings,plus ones, links, diggs, sphinns, or any other easily to measure number you prefer – it doesn’t change the story. Do you see the underlying issue?   The problem is, it’s easy to focus so much on what to measure that we often forget about why we’re measuring it in the first place.

This is quite common in large corporations where departments are siloed and it’s easy for people to focus on their own goals rather than the company’s best interests, but it can happen within smaller companies too – usually when people stop asking “why?”  (Pro Tip: your bonus is nice but your salary pays more – focus on optimizing that.)

Ok, back to the story. Our executive was right. Youtube views did correlate highly with sales, but let’s remember that correlation doesn’t imply causation. (Seriously, SEOs, please understand that point.) Just because two things are correlated doesn’t mean that one causes the other. It’s important to understand why Youtube views correlate with sales. For this, we need to go back to our Marketing 101 sales funnel. It kind of looks like this:

AIDA funnel

Awareness, Interest (or consideration), Desire, Action. There’s probably 200 other versions of the sales funnel but we’ll stick with this one. I want to focus on the top of the funnel. That’s where metrics like Youtube views, Facebook likes, and even many of your advertising campaigns fall. They’re what we call “High Funnel Activities” or “HFAs” for short. The nice thing about HFAs is that they often lead to LFAs (Lower Funnel Activities.) Both HFAs and LFAs are subclasses of KPIs but we’ll cover all of that in Marketing 201 next semester. For those who are still taking Marketing 101, the marketing funnel actually works just like that funnel at the top of your beer bong; the more you fill the top the more that comes out at the bottom. Theoretically, increased awareness leads to increased action – except that didn’t happen in our above example. Why not?

In our example we were so focused on ensuring that the top of the funnel was full that we forgot about the bottom. Instead of pouring more into it, we just stuck a cork in the bottom. The top stayed full and we hit our awareness goals, but only because we trapped customers in the awareness stage and wouldn’t let them leave. In the real world, it’s quite similar to replacing the cash register with a sales pitch. Customers are standing in line with money in their hands and instead of taking it you’re telling them how awesome your product is. Stop talking and let them buy it.

It may sound extreme, but I’ll reiterate that I’ve seen several companies face this exact issue in the past. They didn’t understand why things were correlated so they couldn’t properly use that correlation to their advantage.

Analytics isn’t about compiling numbers. That’s called reporting and it can be automated (or assigned to interns.) Real analytics involve insight. Analysts need to understand not only what is happening but why it’s happening and what we should do about it. That’s called actionable reporting and if you’re interested in that you won’t want to miss the performance SEO metrics session featuring me (@ryanjones) @rhea, @vanessafox, and @statrob at SMX east next month.

In the meantime, here are some tips for succeeding with analytics:

  • Just because something is easy to measure doesn’t mean it should be measured. Make sure you understand why you’re measuring it.
  • Correlation is NOT causation. In fact, optimizing one side of the equation may actually change the correlation.
  • Understand your goals first, THEN pick metrics that line up with those goals. DON’T establish goals based on metrics.
  • Don’t miss the forest behind the trees. Look at all your metrics together to tell the complete story.

About the Author

Ryan Jones

Ryan Jones is a manager of SEO strategy & analytics at SapientNitro where he oversees a regional team providing strategic insights for fortune500 clients and occasionally attempts to take over the world – which he would have already succeeded in doing had it not been for those meddling kids and their dog. The views expressed in this post are Ryan's alone and not necessarily those of his employer or clients. Follow Ryan on twitter at @RyanJones or learn more about him at http://www.ryanmjones.com or on


10 thoughts on “Wait, why are you measuring that?


  • Matthew Egan on said:

    There is an active conflict over data. Different SEOs define value in data differently, and so if you focus on an area of the data that you feel is the most actionable, you’re leaving yourself open for someone to come in and talk “jargoneese” to your client and impress on them that you were tracking the wrong data.

    Inversely, I’ve gone into meetings doing exactly that. Clients have never heard of “non-branded search” vs. just being told their total visits or pageviews.

    There is a lot of education that is needed client side, but as long as the data is a tool of the sales cycle, that education is going to be severely lacking. For every Outspoken Media out there, there’s five Reach Local’s who create data sets and PDF decks that only speak to the product they’re trying to sell and not anything relating to data that might add value to a given client.

    It’s depressing, and I’m honestly not sure what to do about it.

    Scott Stratten said it best, “There are more people selling the dream than there are pursuing it, be careful.”


  • Ryan Jones on said:

    You make some good points. All too common in this industry.

    There are quite a few SEOs out there who pick metrics they can change, then set about to change them and “show” the “impact” they’ve made for the client – even though that “impact” may or may not align with the client’s goals.

    It’s an uphill battle, but I think we’re on the right track toward changing that type of thinking.

    Most of what I do all day involves client education and SEO evangelism. The bigger the client or company one works in, the more important it is to set expectations and ensure the client and SEO team are on the same page.


    • Matthew Egan on said:

      I hope you’re right.

      I’m attending a meeting later in the week where an agency we’re working with on a project told us “well, we can just hide all that in the meta tags” and thought themselves clever. It’s a constant uphill battle and we’re usually the pricks for being outspoken and trying to correct the status quo.

      If the battle gets any more uphill we’ll have to turn to Adam West for tips on scaling tall buildings with only a Batarang and some rope.


  • Todd on said:

    Hi Ryan, good post.

    I too feel as though the industry is so entrenched in the wrong types of metrics and wrong types of reporting. One of the issues is that marketing and analytics is ground-level – on the pulse of the industry. Information trickles upward to CEOs and others within organizations who should be relying on real actionable data using Avinash Kaushnik’s principle of asking “so what?”

    I think CEOs and managers ultimately catch wind of principles that worked years ago, and inflict those painful educational seminars about bounce rate in meetings, while much of the team is much further ahead in understanding what it all really means in relation to KPIs


  • Nick Stamoulis on said:

    Great topic. There are countless tools out there that will provide us with lots of data, but it’s meaningless unless we really know why we are looking at the data and know what we can take away from it. I’ve seen people get so bogged down in data that they aren’t taking the time to take any action based on that data.


  • david k waltz on said:

    Ryan,

    As a finance guy, we are often asked for information and from early-on are taught to find out why people are asking since it is not the information per se that they want, it is the conclusion / confirmation / disconfirmation it provides. Yet often what they are asking for will not serve their underlying reason for asking it, so through understanding the motivation better you are able to provide better data.

    The other thing I caution about (and this is unusual for being a finance guy) is being too focused on metrics. A business organization is a holistic enterprise, and one cannot reduce it to a metric or two, or even a hundred. In my upcoming blog post I note that “one cannot drive a car simply by looking at the instrument panel”.

    Thanks for the perspective!


  • Andy Kuiper on said:

    “…Correlation is NOT causation” so true.
    I don’t know how many times I’ve had to spend a lot of time and energy ‘explaining’ away assumptions per analytics. Nice article; it should be read (twice) by everyone who has a ‘say’ as to how the company website should move forward.
    * p.s. the G+ button doesn’t appear to be working on this post.

    Thanks
    Andy :-)


Leave a Reply

Your email address will not be published.

*


*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Comments links could be nofollow free.