min to LISTEN
January 9, 2018

How to Turn Website Data into Insights You Can Actually Use

Adam Greco
Adam Greco
Senior Partner
Analytics Demystified

My guest in this episode is Adam Greco, a web analytics expert who has worked for companies such as Omniture and Salesforce.

Most people are easily overwhelmed with data analytics and don’t know where to start. Many organizations also function in a vacuum with each department using their own data and not sharing their findings across the team.

As users we don’t want reports, we want insights to help make decisions. This episode we’re going to talk about how to turn that data into insights you can actually use, how to spot problems on your website and sales funnel, how to score your leads, and how to know what to focus on.

Adam is a prominent member of the web analytics community and he gives us a practical guide to analytics you can start implementing in your business or project right away.

Listen to this episode:


It's the antidote to marketing bullshit.

Receive a free, 8-lesson video course + a super practical, no-bullshit essay in your inbox every Tuesday.

Join 4,003 mavericks getting my emails,
but who's counting, amirite?!

"Anything with Louis always blows my mind."

Zineb Layachi
CMO in Residence

"Louis is a genius. Sharp and super useful insights."

Maja Voje
Marketing Consultant

"You're literally the only marketer I can stomach."

Braeden Mitchell
Security Engineer

We covered:

  • The actual goal of web analytics
  • The “there is no spoon” exercise: why do you need a website anyway?
  • Looking for problems in an ecommerce funnel and convincing your managers to act
  • Best segmentation methods to isolate a problem: score-based, organic vs. paid traffic
  • User feedback beyond web analytics
  • Testing, personalization and getting ROI out of web analytics
  • Salesforce case study: ethics versus data
  • Adam’s advice to young marketers: technology and programming languages to learn
  • Adam’s recommended resources


Full transcript:

Louis: Adam, what a pleasure to have you on the show. Here’s a typical situation. A company has a website and they have Google Analytics set up. They may have Hotjar or something along those line set up on their website as well. If it’s a company that’s quite big and sell an advance product, they might have other stack on top of it, maybe Adobe products or anything like this for analytics. But regardless of that, it seems like the normal process is to collect data, to re button it, and then that’s pretty much it. Why is that?

Adam: I think a lot of people don’t really understand the goal of web analytics and sometimes, it’s because it’s being driven by tech people versus business people. But at the end of the day, the real over arching goal of web analytics is to provide you directional guidance on what’s happening on your website so that you can make better decisions and then you can improve your conversion rates, whatever that is. If you’re an ecommerce site, that’s obvious. If you’re not an ecommerce site, that might be a little bit trickier to figure out.

But I think, people start too often with just here’s a bunch of data that I have on the page and here’s all of this information so I’m just going to collect this data and hope that maybe someone is going to use it. That’s like looking for a needle in a haystack and not something I recommend.

Louis: Here is the problem we’re going to try to address in this episode, is that most people don’t understand how to make sense out of that data. As you said, they are in a vacuum. What turns out is users don’t want reports. They want insights, as you said. They want insights on how to increase conversion rates. Let’s try to solve this problem right now. What I’m going to challenge you to do is really to try to come up with a step by step methodology. Something that people can take away and apply in their business, regardless of their size.

We’ll answer a question how to use data strategically to get insights, just basic solution we’re going to try to layout together right now. What will be the step one of this process? Of this scenario?

Adam: Okay. The way I would recommend is if you work at a company and you’re not sure if you’re really getting the most out of web analytics, step one is what I call, for those of you who’ve seen The Matrix, I call it the “There is no spoon exercise.” What I recommend people do is pretend that your website just disappeared. It was nuked. The server went down and it doesn’t exist.

You have to go to your boss and you’ve got to justify a huge amount of money. Let’s say it’s 10 million euros to build a website. Imagine your boss is kind of like my father in law who’s an old [00:04:49] who thinks the internet is just a fad and you have to convince this boss to say, “We’re going to spend $10 million.” And they say, “Why do we need a website anyway?”

If you take a step back, that forces you to say why does our website exist at all? Why do we even have a website? If you then say well, our website allows—let’s say you’re a B2B company like Hotjar or where I used to work, salesforce.com and you have to go to David at Hotjar and say, “We need a website.” He would say, “Well, what is it going to do for us?” And you’d say, “Well, one thing. We could actually show people our products 24 hours a day, 7 days a week and we don’t have to always be awake or have people around the clock.” David says, “Okay, that’s cool. Let’s definitely do that. How are you going to prove to me that our website is actually teaching people about our products 24/7?”

You would then have a KPI and you would say, “Well, we’re going to have a KPI called product views and then we could break it down by which features or which products they’re looking at.” And so now, you’ve got your first business case and your first KPI that you can then measure to see how the website is performing and you just repeat that for all of the things.

I have found when I work with companies large and small that their implementation has been handed down over and over from people to people and they sometimes lose sight of the forest for the trees and if you force them to use the spoon exercise to say why does our website begin? You can get back to why you originally even paid for a web analytics tool to begin with.

Louis: I really hope that your step dad doesn’t listen to podcasts either.

Adam: No way.

Louis: Okay. So we can talk about him. It’s a good advocate for you I suppose. Right. I very much like this step one. I’m not saying that just for the sake of it but I think this is probably one of the best way to explain this concept of picking the right KPIs and knowing why you’re doing what you’re doing ever. I don’t think I’ve listened to a better explanation than this for this particular subject so it’s really nice to have this analogy.

Let’s say now we know why we have a website and we know what type of numbers we want to track. What is the next step?

Adam: Now, the next step, I won’t get too technical here but obviously, with web analytics, you’ve got to work with your developers and you need to basically put some code on your site where you’re going to start collecting data and let’s imagine that’s happened for a couple of weeks. Now, what you want to do is you need to dig into that data and say what trends do I see? Where do I see people dropping off of keypads? Where are my metrics pretty consistent? Are they not consistent?

That’s when you start doing your actual analysis. A lot of that will have to do with segmentation, where you want to look at how do first time visitors act differently than returning visitors. Basically, the goal of your second phase is to say is there anything that I can find that is really interesting, that I want to dig into more?

You’re not going to solve it but you’re going to basically just look for, think of it as looking for fires that maybe you could start putting out. The key with web analytics, this is where some people misunderstand web analytics. Web analytics is not going to tell you why your website sucks or why you’re having problems. It is there to be like an early warning detection system to say hey, of all the places you can look, look over here and look over here because the data is suggesting there may be a problem. That’s all you want to do out of step two, is find out where your problems are and then I’ll go into later when we get to the further step, how do you actually diagnose it. But that’s step two.

Louis: Alright. Let’s dig into the problem itself from step two. You mentioned a few ways to quickly know whether there is a problem in the first place. What are the typical ways? Let’s take a simple example. You have Google Analytics set up on an ecommerce website, probably one of the simplest thing to set up, and obviously, the end goal is a purchase. You might have tracking on each step of the funnel. What is the usual exercise or things you look at to see whether there’s a problem?

Adam: It’s different for every site. But if we use ecommerce as an example, everyone has different approaches of how they go about it. I can tell you how I do it. I always like to start at the money. That’s what people care about. Where am I making money? Where am I losing money? For example, I will start at the purchase and that pages around the purchase and the funnel and say of all the people that are not purchasing, how many are making it to the step right before the purchase, which is usually the check out.

And there, right even in that one thing, you could probably spend a couple of months just trying to figure out where are people dropping off for certain products, for other products they’re doing okay. Where are the marketing channels have an impact on the conversion? Are people from SEO converting from the checkout to the purchase more often than people from paid search or display?

But then, you take it a little further back and say of all the people who add products to cart, how often are they making it to checkout and then making it to purchase? And so, I think of it as concentric circles with your dollar signs or euro signs in the middle. As you branch out in concentric circles, the amount of impact you can have decreases because you want to really focus on the people who are telling you they want your product, they’ve made it all the way to checkout but they’re not converting and then do some AB testing. Maybe you try some different kind of conversion funnels and so on.

But I like to focus really on the basics there and then work my way backwards. A lot of clients, what’s funny if you use a normal Google Analytics implementation, just as an example of how people screw this up, you will normally track how many people put a product in the cart and they’ll track it with the product ID and a cart ad and then they can see which products make it to checkout and purchase.

But one of the things that not many people do is even track something as simple as how much money is being added to the cart by product. And what’s really cool is I’ve learned over the years that if you go to an executive and you say, “Our cart add to purchase ratio is only 2%.” They’ll say, “Wow. That’s okay. We need to make that better.” But if you go to them and say, “Hey, we actually had 10 million euros added to the cart and we only sold 1 million.” They will freak out. They’ll say, “Wait a minute, we’re leaving 9 million on the table here?” Even though that may be the same 2% conversion rate you just told them, they will snap into action when they think they’re missing out on 9 million euros versus a stupid percentage that doesn’t really mean anything out of context.

Louis: This is a great tip. Let’s do a parenthesis here about how to convince your boss or your managers to act. This is pretty much exactly what you said. Instead of using numbers that are very complicated and even percentage, people don’t really understand what percentage is if you dig into it in more detail and it’s always better to talk about as you said, monetary value, actual amount of money or stories.

What’s also interesting is sometimes, we just isolate a particular case of a particular person, a natural person who went through the process and struggled and tell this story, tell her story and that will also help people to really understand that there’s an issue. I very much like this approach. Let’s take a step back. You mentioned a few things that seems obviously trivial to you because you’re the expert, I’m not.

But I want to dig into something specific. You said you focus on the end goal. For an ecommerce site, it would be the portrait. For a SaaS website, it might very well be the first subscription, the first time they subscribe and then you take a step back, just a step before and you try to see how many people have actually completed this step before moving on.

And then you mentioned segmentation. You said segmenting by traffic type, segmenting by profiles. What are the typical segmentation that you would recommend companies to use to isolate a problem and to see if there’s a problem?

Adam: The common ones that I see people use are the ones you just mentioned. How are they finding us? Is it a paid or is more of an earned visit to the site? How many visits have they had so far? How long is it taking them to convert? Is this their first time? Is it their fifth time? I also think that nowadays, because people are working with multiple devices, another segmentation might be are they on an iPad? Are they on a phone? Because maybe you don’t have as good of an experience on a mobile device as you do on a desktop site. So that’s another thing that people would look at.

And then there’s a whole world of segmentation around do you know who these people are? Some websites, they actually know people. If you’re a banking site, you can’t really do much on the site without having logged in. So if you know who the person is, that gives you a whole world of CRM of data you can integrate with your web analytics tool to segment on gender, age. But for most people, we don’t have that information and that’s going to get even worse because the GDPR stuff that’s happening especially in Europe.

A lot of it is more anonymous things. But one thing that I will tell you that’s a little more advanced, that I love to help clients do, is segment on what I call visitor engagement. The way this works is that if you have people coming to your site multiple times, you may not know who they are, but you can actually score them based on the actions that they’ve had.

Let’s just imagine, for example I’ll use Hotjar. Someone comes to the Hotjar website for the first time and they look at a product. You might say, okay they’ve looked at a couple of the Hotjar features so they have a really high score for heat maps and recordings but they have a really low score for some of the other features of the product like polls.

In that case, you can now look at people and narrow down your conversion of people starting a trial based on which things they’ve looked at. Their score, if they come multiple times, you can keep adding to their score and you could even make it numeric. You can say, “This person has a score of 75 because they’ve been here multiple times. They’ve clicked on our pricing page, which means they’re really interested so let’s jack up that score by 10, 15 points.”

And then, at the time that they submit a trial, you could actually document what was their website score and you could segment to see if people with different scores convert at different rates. A perfect example of this is kind of fun from my time at salesforce.com. As you probably know, sales people hate marketers more than anybody. Oftentimes, in the B2B world, the sales people say, “We do all the work and the marketers give us crappy leads.” And then the marketers say, “We have great leads and the sales people don’t know how to sell it.”

At salesforce.com, we had that dynamic where sales people hated marketing and we did this type of scoring. We said every time a lead form is submitted on salesforce.com, we passed their website score into our CRM and then I looked at two weeks of data from our sales team and figured out when the sales people called prospects, how often did they get hung up on and said, “You’re spamming me. Don’t ever call me again.” I looked at the scores there. It turned out that anyone that had a score of under 50 basically got hung up on. And if they were over 50, they were much more receptive.

What we were able to learn there is that if we stop calling people right away and let them nurture a little bit, then they were much more open to a conversation. We used score based segmentation to hire way fewer sales people and we were able to get the same number of deals with a lot less sales people by just looking at their website engagement.

Louis: This is a fantastic tip. I know it’s quite advance but this is the type of stuff that you will not really think of on your own. That sounds really good and it makes sense. Let’s say I go to supermarkets and there is this annoying salespeople trying to sell you stuff. You go there and this person says, “Hey, if you want this holiday in Spain.” And they ask you straight away to purchase from it, you’re way, way, way less likely to do so than if you receive a brochure in the post or emails or if you have previous relationships with the company and yes, the last touch point is with this salesperson that you actually know in the supermarket. You wouldn’t do that in real life yet we do that online quite a lot. We expect people to just be convinced straight away and be like, “Oh my God, this website is so good. I want to have a call with somebody straight away.” And buy a product. It never happens.

Adam: A good analogy I’ll give you that I use sometimes is if you’re single and you walk into a bar, you don’t normally walk up to a person and say, “Hi, my name’s Adam. My place is right around the corner. Let’s go.” It just doesn’t work that way. You have to get to know someone. You talk and you date. Marketing is the same way but sometimes, marketers, they try to close the deal way too early.

Louis: Absolutely. Treat people the way you’d like to be treated, basically. I want to dig into more basic stuff. You mentioned segmenting your traffic type. Let’s say we want to compare organic traffic performance versus paid traffic. How would you typically spot a problem whereby for example, paid traffic is not performing well compared to organic? What’s the typical way to see that?

Adam: Mostly, what you would do is open up your key reports in your analytics tool and do two things. One is you would create path reports that show you the flows the people are taking for group A, which are your paid and then group B are your non paid, and then look to see where they’re dropping off and see what’s going on there. And then your other report is you would look at your standard conversion reports.

For example, if I’m an ecommerce site, I might look at what is my look to book ratio, which is defined as how many people looked at my product and bought it? And then be able to break that down by paid versus organic. And then within paid and organic, you can then break it down even more granularly to see by keyword, by campaign code and so on. You might see that people from Twitter tend to do better than people that come from Facebook and so on.

It’s just a matter of looking at your conversions and splitting it out. But this gets to the other point that we mentioned earlier, is that your web analytics tool is really just there to tell you where you might have problems but I actually think that very few people get to what I call phase three, which is the most important part, which is why, why are they not converting? Why are we having problems on our website? And that’s where you have to get out of Google Analytics and you have to get out of Adobe Analytics and move into other tools that can give you the bigger picture. That’s where the Clicktales, Hotjars of the world are able to add more color to the situation because you can take the problem you’ve identified and then you can ask people why are they having this problem.

For example, if you’re salesforce.com and you see that you have 20% of the people who view a form, complete it, but why are the other 80% not completing a form? Well, you could look at web analytics data all day long. It’s not going to really help you. Or you could actually ask people who are on the form and don’t complete it and say, “Can you explain to me why you didn’t fill out this form?” And get some real answers from real people. That’s how you can make a huge step change in your conversion rates versus looking at numbers all day long.

Louis: Let’s take a typical example. Let’s say that I’ve noticed that people from organic search are actually converting twice as much as people coming from AdWords and paid traffic. Let’s say we have a simple ecommerce store and the thing that we are tracking is the purchase. Those people convert twice as much. How do you find out why those paid traffic visitors are not converting? In practical terms.

Adam: That’s hard. There’s not an easy answer. You have to look at all of the attributes that you have combined. The visit number, how long they’ve been there, what content on the website that they’ve read and does that have an impact on conversion. There’s no great silver bullet. I think one of the things that people have been waiting for in the web analytics space for a long time is the addition of machine learning and artificial intelligence where you can collect a lot of data and actually have your analytics tools tell you the answer and say, “It looks like this is the problem.”

Adobe is trying to get into this area. A lot of other companies are playing in this area, where they’re basically saying export all of your web analytics data and let machine learning go crazy on it and point out things. There’s one funny story that happened here in the US, where it’s a famous story with Target, a retailer here, where there was a teenage girl that was surfing on Target’s website and looking at a lot of things. In the mail, she got some sort of an advertisement, I don’t know, I can’t remember all the details, selling stuff for people who are pregnant with babies. Her father got this in the mail and got really upset and started yelling at Target and said, “Why are you sending stuff to my teenage daughter?” And it turned out that his daughter was pregnant and he didn’t actually know it. But the machine learning that Target was using has actually figured out by the things that she was looking in on the website that she was pregnant and they knew that she was pregnant before her father did.

That’s the power of machine learning that people are looking forward to, even though it’s a little bit scary at times.

Louis: I did read this story. I don’t remember from which book but I remember reading this. It’s pretty intense and I think we’re going to have to find ways to prevent this type of intelligence to go too much into the intelligent type of things. I think as you said, with GDPR and the privacy concerns that is there, personalization needs to find a common ground. Somewhere where we don’t stalk people but we still provide value. We’ll see where that goes.

Adam: Yup.

Louis: Anyway, in a perfect world, we’ll have an analytics tool that tells us what’s wrong. How do you get to the why?

Adam: I think the why is more about asking questions to people. That could be done through the old fashion ways of focus groups or nowadays, there’s so many tools out there. You’re familiar with Hotjar and others where you can actually watch people using your website almost like a TiVo for your website. Those tools, they’re really powerful but I also think that they can be problematic if you don’t combine them with your web analytics tool. For example, what I really like to do is say, “I’ve identified a problem and here is a segment of visitors.” Let’s say it’s people who add over $300 to the shopping cart in this month but didn’t purchase. I don’t want to really go look at Hotjar and look at 1,000 recordings. That would take forever.

But if I could actually say, “You know what, there are actually only 50 cases where this happened this month but that’s $300 is a lot for our site so I would like to watch a bunch of just those recordings.” That’s where it’s really cool where you can combine the best of both worlds to say the web analytics tool has found the problem. Now, we’re going to go watch the recordings and see if we could identify what the issue is.

Now, watching a bunch of recordings can sometimes tell you, sometimes it can’t. You might have to do a little more that says okay, we think the problem might be this so now, we’re going to pop up to maybe 10% or 20% of our users and say, “Is this A, B, C or D? What are the reasons why you’re not purchasing this product? What might be going on?” You can survey them.

A lot of people don’t like getting pop ups but I’ve actually found that if you do it the right way, you can’t be, as you say, on the dark side of marketing. There’s a way you could do it and still be on the light side by just not abusing it and really being focused on the questions you’re trying to ask them when you ask them.

Louis: I like this idea because we tend to really drown into the sea of data out there. If you watch session recordings and if you look at heat maps and all of that, if you don’t really know why you’re looking at this data, it’s very difficult to make sense. Start with the what, identify the biggest problems and then understand why it’s going on very much like that.

Is there any step after this step that you would recommend listeners to use?

Adam: Yeah. I think when you get more advanced with analytics, I would say the next step is to really go deep on testing and personalization. There are great tools. Google has tools, Adobe has tools. Most people are using Optimizely. Optimizely is a great testing tool where you can then say, “We know we have a problem in this general area. We’ve gotten some feedback from people but now, we want to test a different version of this page or we want to actually test a whole different conversion funnel.

For example, I had one retail client who said we want to understand if guest checkout is good or bad and maybe if we add PayPal, if that will really help our conversion rates. What they did is they did some AB testing of different flows and they would say if we ask people for less information, do we get more people through? I think testing is where a lot of this gets paid off in the end.

The analogy that I give, this actually, I steal this a little bit from one of your past guest Seth Godin. Seth Godin has this great expression where he says in marketing, you should always strive to be a thermostat instead of a thermometer. The way that I apply that to web analytics is if you’re just looking at data and just looking at reports, you’re basically a thermometer. You’re telling the temperature of what happened today, what happened yesterday. But to really be a good marketer, you want to be a thermostat, where you’re basically pulling different levers to change the future.

And so, if you do analytics, you can understand what’s happening but until you really start doing testing and say, “I’m going to change something on this site and then I’m going to see the financial impact of that,” you’re really just a boring thermometer. I love when people change things on the site because of data. The scary truth that not many vendors want to talk about in the analytics space is that you actually get zero ROI from web analytics, both the tool, the people, all the time you spend, if you never make a change on your website based on the data. If you just report on stuff, then basically you’ve just made web analytics an expense of your company. It’s just like a cellphone. It’s just like a computer, except you can’t really depreciate it as much.

But if you actually use the data to make a significant bunch of changes on your website and then you could see the impact of that, now you have ROI. I push my best clients to be able to say, “Even if we pay $1 million a year to Adobe or Google for our analytics tool, we can prove that we’ve actually generated $2 million or $3 million worth of savings or incremental revenue by using those tools.” But I am shocked by how many people throw up their arms and say, “You know what, web analytics is an expense and it’s just something we have to do.” That’s a shame. That’s not how it should be.

Louis: It’s a great way to end this section of the interview. I love this saying by Seth Godin. I actually haven’t heard it before. It’s quite nice to end up this step by step. Thank you so much for taking the time to go through it in detail. I think it’s going to be really helpful.

Adam: I think you proved that you are quite an influencer. I’m only messing with you. You’re a very well known person in the web analytics community and beyond that, in digital marketing community. As you said, you’re a web analytics consultant and you have years and years of experience for companies like salesforce.com. You’re also in the board of directors of the Digital Analytics Association and you now serve as their treasurer so I’m sure you can tell us a lot about good case studies and very, very successful clients that you helped to grow their business.

What I’m interested in hearing from you today is actually a fuck up. I’d like to hear from you, you don’t have to tell the name of the clients, one of the biggest fuck up you’ve made as a consultant and what happened and why did it happen and what did you learn from it.

Adam: Okay. I would say it’s kind of a borderline fuck up. But when I was at salesforce.com, we had an age old debate. The debate was do we put forms in front of everything that we do on our website? In order to get a demo of the product, you have to fill out a form. If you want to do a free trial, you have to fill out a form. The number one feedback that we got from salesforce.com, net promoter score, everything, was we like you but we hate that we have to fill out a form to get everything and that’s so old school.

Me and a couple of other people were really big on this idea of let’s get rid of the forms. We think this would be a better experience and that means that people will use our product in a free trial. They’ll do demos. They’ll tell their friends and in the long run, we’ll be better. But that was a really risky endeavour.  How do we test that?

We went to a very elaborate test to basically say we’re going to take 10% of the people who come to our website, we’re going to cookie them and we’re going to let them not see a form in order to do a demo and a free trial. We documented those people and then waited a couple of months to see what happened. We were so hopeful and so optimistic and in the end the people who filled out the forms converted so much more. The sales people were able to call them and harass them to the point where they actually bought salesforce.com and the people who didn’t see the form, a lot of them, we couldn’t track in the next three to six months that they had comeback and proved that they had eventually joined.

Our team was really hopeful that we can make our website experience better by not having to have these forms. But at the end of the day, we were wrong. The truth is that you can’t be a believer in data and believer in web analytics and say, well, the data doesn’t show this but I still think this is right so let’s still get rid of the forms. I had to kind of eat a little bit of humble pie and go back and say you know what, you live by data, you die by data. In this case, we had to stick with the forms.

I still, in the back of my mind, believe, that if you look at a longer horizon of a couple of years, maybe we would’ve been better to get rid of the forms and maybe we would’ve gotten more word of mouth and it would’ve been more like some of the premium models like Hotjar and others use. But it was a really difficult thing to prove in analytics and it frustrated me to no end.

That’s the other lesson that I learned. You can’t get emotional about your decisions. Sometimes, you have to go with what the data is telling you and it may not go your way.

Louis: Thanks for sharing this. If that’s your biggest fuck up, then you’ve had a pretty good career so far. The point here that you’re making is a very interesting one and I’d like to start on it. I haven’t prepared for that but we can improvise as usual. I believe that you can’t measure everything in marketing and this is I think one of the mistakes that we do if we try to measure everything. One of my past guest DHH for Basecamp talked about this quite a lot. He was talking about retargeting and the fact that people were getting so caught up in the fact that oh, it takes only $5.67 on average to get this new client on board. He was making a point that retargeting is only the last touch but there are a lot of touches that you can necessarily measure well and therefore, you can’t just say that retargeting was responsible for creating this client.

In your example, I might challenge what you’re saying, by saying that it’s not because you can’t measure it, that it didn’t improve things. As you said, you were in a situation where you had to prove it to implement it. You weren’t able to track those people that actually didn’t fill the form and maybe were happier about the experience, therefore talked to more people about it, therefore more people came by to the website and converted in return. This is the crazy thing about marketing. I believe that you do have sometimes to take ethical decisions or decisions that are based on what you believe in and go against your numbers when you feel that this is the right thing to do.

Adam: I completely agree. It’s a fine line because I also think sometimes people are too focused on data and it almost can stifle creativity. If you think about some of the best things that marketers have ever done are things that they’ve taken a leap of faith and they’ve said, “You know what, I think this is just the way to do it.” For example, I think we should be in content marketing and we should share a lot of great things about our product.

Like for example, I’m in a consulting role right now. I have probably blogged about everything I’ve ever done related to my field, Adobe Analytics and people will say to me, oftentimes, you’re crazy. Why are you telling everyone? Because now, they’re not going to pay you as a consultant. But I’ve learned that the more I give away, kind of like Rand talked about, the more I give away, the more people trust me and the more they actually come to me. But I think that step changes any process oftentimes happens in those moments where you’re in the shower and you come up with this crazy idea. But I think you could still use data to test out those crazy ideas that you have.

But in this situation at Salesforce, I felt like in some cases, we may have been too tied to the data and it wasn’t my group. It was the people who we had to convince that maybe if they would’ve let it go for longer, if they would’ve just seen the bigger picture, but they have a quarterly number to hit and they see that people aren’t filling out forms, it’s kind of hard to argue with that. That’s the good and the bad of analytics. I think Jason Fried, who I’m a big fan of, he often says, I don’t believe in web analytics at all. He’s like, “Listen, I just think that I know what is best because I know what I like and so on.”

There are areas where I agree with that but there are areas where if you don’t look at the data, you may not be representative of your userbase. I think there’s a line and I think every company has to balance that line.

Louis: That’s true. Jason Fried is the co founder of Basecamp and they do measure things pretty intensely. In the episode with DHH, we actually talked about it for a while, about numbers. They do measure stuff. As you said, the line is blurry. It’s not either/or. You can’t just believe in it or not believe in it. You just have to pick your battle sometimes.

I firmly believe that marketing is really emotionally driven. I firmly believe that it’s based on as you said, giving value in order to be trusted and you can’t really put a number on trust. Sometimes, it’s just a matter of doing the right thing over and over and over again and then it pays off in the long term.

If companies really try to hit their quarterly target all the time, they will miss the bigger picture. The fact that a brand takes years for you to build and it will just take time so you have to make peace with it, I suppose.

Adam: Yup. Exactly.

Louis: Moving onto marketing in general. I think it’s a good transition. Why do you think marketers have such a bad reputation in general?

Adam: I can’t speak to all the reasons why marketers have bad reputation, what I can tell you on this topic is I think analytics marketers tend to get a bad reputation for a couple of reasons. One, they don’t really listen to what their stakeholders want to know and what are the questions they want to answer. I also think that there’s a huge gap in our industry with what I call data quality, where you’re tracking things in a web analytics tool but you’re not actually checking to see if the data that you have is right.

You talked about attribution issues. Is it first touch? Is it last touch? What if they’re using multiple devices? Some analysts will say, “No, this is exactly what’s happening.” When they don’t really know because people are coming from different devices and they can’t do that.

I also think that many digital marketers who I work with, who are in the analytics space, they spend 90% of their time stuck in their cubicles. I call these people Miltons from the Office Space movie, where they’re sitting in their cubicle crunching numbers all day and they don’t actually go out and find out what people want to know and then they just come back with these boring reports that is just the same report you saw last month with new numbers.

I think that’s what gives digital marketers a bad reputation. They say well here’s this number guy. It’s kind of the same thing that happened in sports with Moneyball if you’ve seen the Moneyball movie, how all of the traditional scouts hated the numbers people because they’re like, “Oh, these are just nerdy people who are coming here to tell numbers.”

But once they actually got out and talked to the baseball players and started explaining how the numbers helped them and how they applied to them, suddenly, they got on board. I’ve listened to many of your podcasts and I agree with most of the things that people are saying in terms of why people hate marketers in general. But I can only really speak to my domain, which is why do people hate digital analytics people when I interview companies.

Louis: Yeah. That’s what I was asking so it’s a great answer. To add to your point, I guess what happens when you focus on numbers too much and really forget about the people behind is that what’s called a statistical numbing. It’s like 10,000 hits on a website doesn’t seem like much if you’re just used to looking at it everyday. But there are actually 10,000 people going through your website if you exclude bots and all the stuff like this happening at the minute.

Imagine 10,000 people in a stadium in front of you, that’s exactly how many people are going through your site everyday. I do believe that it changes your perspective on things. Spending time with those people face to face or on Skype or on other channels instead of just looking at numbers will definitely add this new element to your work as a web analyst or as a digital marketer.

Adam: Exactly. I’ll tell you a funny story that plays right into that. When I was at salesforce.com, there was a new product that our CEO, Marc Benioff, who’s a billionaire, very impressive guy, came up with called Salesforce Chatter. It was basically when Facebook was really coming of age, the idea was it was Facebook for your company. You could post stuff to your CRM as your feed of what’s happening. They were really gung-ho on it and it was going to be like the next big product at Salesforce.

They came up with a homepage design that literally was a flash animation of these chattering teeth that I wish I had a visual to show your listeners. But it was a really scary flash thing that was just chattering and it was this idea of chatter. We all looked at it and we were like, “Wow, these teeth are really ugly.” And we’re like, “This is not going to do well.”

We had this really boring homepage that just showed our products but then suddenly got taken over with this animation and our bounce rates shot up. Our conversion shot down. Like you were saying, these were real people that were coming to our website and being freaked out by these chattering teeth. We tried to use numbers to convince a bunch of the executives at the company, “Hey, we got to get this thing off at our homepage.” And they were like, “No. Listen, this is Marc’s baby. We’re not taking this off the homepage.”

What we actually did is I used a product called fivesecondtest.com and took a screenshot of it and went out to a group of people online and took away all of our branding and said, “What does this company do?” The responses we got back were awesome. They were so comical. Comedy store, dental office, all these things that had nothing to  do with CRM. I went back to my boss and my boss’ boss and said, “Listen, I love the chattering teeth—because I was a suck up and didn’t want to get in trouble with Marc, I said—but our customers seemed to be confused of what we do here at salesforce.com once we put these chattering teeth on the homepage.

It got them to empathize with our users and said, “Is this really what you want to convey to them?” Because these are real people who are confused. Like you said earlier, the bounce rate didn’t seem to make an emotional charge with them. But when they saw the phrases, the verbatim responses from these people, they realized you know what, we do need to tone this down a little bit and take that off the home page.

Louis: That’s a great story, especially from inside Salesforce. It further proves the point that if you really want to convince people, especially the C-Suite or your boss, much easier to prove your point with data and actual people saying something. One of the best ways to really record videos of people’s reactions or what they say and you only need five or six people. Even sometimes, just one to make people change their mind.

As you said before, you can show them numbers for ages before they change their mind but one person or two people can really change people’s perspective. Thanks for sharing this story. What do you think marketers should learn today that will help them in the next 10 years, 20 years, or 50 years?

Adam: I think marketers of today have to understand data and know when to use it and all the different tools that are at their disposal. It used to be in the old days that you didn’t have to know anything about technology to be a marketer. I think that ship has sailed. I actually tell all the young people who I know who are majoring in business in college, I say, “No matter what you do, you cannot graduate without knowing a little bit about technology. You need to know whether it’s JavaScript or some sort of programming because marketing is becoming a technology now and you have to be able to speak to developers. You’ve got to be able to quantify what’s happening and you’ve got to be able to see whether it’s the JavaScript code.”

Nowadays, I actually tell marketers they need to focus on learning things like R and Python because you’re going to then have to know how to do a programming language to actually query the data. It’s not always going to just be nice and neat in this little Google Analytics or Adobe Analytics. You might just have to query raw data stores and join data together.

If I were a young marketer now, even though it doesn’t sound intuitive, I would be learning a little bit of programming. I’d be learning a little bit about SQL and I’d be learning about APIs, how do I pull data from different sources and merge it all together and then maybe the tools of the future are going to be more like the tableaus of the world where you’re basically joining data to try to figure out what’s happening across different data stores.

But just knowing all these tools, it’s intimidating and you’ve got to pick the right ones and know enough to be able to be I call a technical or a digital marketer.

Louis: What are the top three resources you would then recommend for people to learn about this concept you mentioned?

Adam: I think that there’s a ton of free resources out there for R and Python. One resource if you are in the analytics space, there is a public Slack group that is called #measure Slack. Basically, there are about 4,000 web analysts that are in there. If you Google it, there’s a web Google form that you fill out and then some guys who I know will actually look at your form and determine if you’re actually in the industry and not just a spammer and then you will get an ID to Slack and you will now see channels for statistics, Adobe Analytics, Google Analytics, data visualization. You can post questions and have a nice community forum of people to meet and you can communicate one on one. That would be one.

Obviously, as a member of the Digital Analytics Association, I have to put a little plug in. We do have a lot of training programs and if you’re a member, which is only like $200 a year, you can basically go to webinars. We’ve actually created this thing at the DAA, which is cool, we call the Analytics Cookbook, where I ask all of my friends to go out and document one web analysis they’ve done in the past and share it. We now have 120 web analyses that you can go in and look at and we’re always adding more every week so that if you’re at a company and you’re just not sure where to start, you can go look at these recipes and go do that to your company.

And then programming, obviously, there’s lots, Code Academy and all those places to go learn R, Python and JavaScript.

Louis: I like this cookbook idea. I need to work on that. I think I have ideas to apply it in my own business.

Right. Adam, you’ve been amazing. Where can people connect with you and learn more from you?

Adam: I’m on LinkedIn, just Adam Greco. I work at Analytics Demystified. I have a blog on our website but I’ll warn you most of my blog posts are really only relevant to people who use the Adobe Analytics stack just because I was one of the founders of Omniture, the company that made that so I tend to focus on that. I also speak at a lot of web analytics conferences so you’ll usually see me around doing trainings or speaking out there.

Thank you so much for having me. It’s been fun.

Louis: You’re very welcome. Thank you once again.