You’re about to hear from a renowned analytics expert. But this episode is about so much more than analytics. It’s about being more deliberate, an essentialist even in your marketing.
Avinash Kaushik is the author of two best-selling books, Web Analytics: An Hour A Day and Web Analytics 2.0. He was at Google for 16 years and recently left to take on the global Chief Strategy Officer role at Croud, a leading full-service marketing agency. He has a popular blog called Occam’s Razor and a newsletter called The Marketing < > Analytics Intersect.
Avinash is an authoritative voice on how marketers, executives, teams, and industry leaders can leverage data to fundamentally reinvent their digital existence. He’s been recognized as the Statistical Advocate of the Year by the American Statistical Association, Most Influential Industry Contributor by the Web Analytics Association, and recipient of the Founder’s Award by Google.
To say Avinash is brilliant would be a vast understatement. I loved this conversation, and I’m sure you will too!
And now, on with the show!
In This Episode
- [02:24] – Avinash describes his journey to get into analytics.
- [08:12] – Stephan differentiates being deliberate from being discerning.
- [10:27] – Avinash emphasizes the importance of being open to new data and making data-driven decisions.
- [14:26] – Avinash elaborates on how he allocates 15-20% of his budget on “insane” things. He mentions risk-taking and the importance of measuring your brand awareness through KPIs.
- [22:18] – What are a company’s typical Google Analytics setup and the implementation challenges?
- [29:38] – Avinash discusses more on Google Analytics, mentioning his take on Google Tag Manager and other plugins.
- [33:19] – Avinash gives examples of micro-conversions from business to business to consumers.
- [37:18] – How to track micro-conversions and set the goal value.
- [43:13] – Avinash elaborates on the importance of SEO in your business.
Thank you so much for coming on the show, Avinash.
Thank you. I’m very excited.
Let’s first talk about analytics and how you got into that as a specialty. What pivotal events, chances, occurrences, or synchronicities happened that brought you into the analytics world?
I’m a mechanical engineer. I did my bachelor’s in India. Then I came to the US for an MBA as a poor student. I got an assistantship with the Office of Career Services in my second year at MBA school, where I had to organize all the students’ data into a database to understand their career paths.
I built this thing as a little project, started creating some reports for them, and I realized it was fun. You can see different patterns in students’ careers and what makes them successful. That was the first trigger—playing with data to help make our Office of Career Services at Ohio State much smarter.
When I graduated from my MBA program, my first job was at Silicon Graphics, figuring out how to help their marketing teams make better decisions. The first tool I built was called MYOB (Mine Your Own Business). It was just a web-based front end on top of a Sybase database that would allow anybody to run queries versus using a client. At the time, people were still using business objects and the heavy software we had to install. So instead, we did everything using the web.
It just started the ball rolling to the extent that from there, to DirecTV, to Intuit, to Google. I’ve always been in a space where I tell my mom that she doesn’t understand everything I do and that I help people make better decisions that have been on the journey since that little assistantship at the Office of Career Services at Ohio State University.
That’s awesome. Pattern recognition is really important. It sounds like you’re very good at that. But even more important is pattern utilization. That’s where you come in with helping people make decisions based on data. It’s data-driven decision-making. It’s finding which data is actionable and discarding the rest or ignoring it because it overwhelms you in inaction.
I agree with you. I teach it at Stanford and Berkeley. I tell the students that early on in my career, my big quest as a person who lived in the analytics space was to collect as much data as possible because having more data meant making better decisions. This is the 1998–2000 timeframe.
With the web arriving and my career shifting more centrally towards digital analytics starting in 2003, I learned quickly that what helps you make better decisions is knowing what data to ignore because we had so much data. They’re just flowing through every corner that is possible. Over the last decade, this has only morphed.
If you and I worked on a digital or multi-channel business, we would easily use 25 tools. We’ve got access to every data. It is more about figuring out what is essential, what the purpose is, what the KPIs are, and how much you should ignore so you’re not in the business of data puking. You’re in the business of making wiser decisions like analysis.
These days, I spend most of my time figuring out what to delete so that we can focus on what’s essential rather than collecting more and more data. Yet, many people still are like, “I would like Google Analytics, BigQuery, SimilarWeb, Yoast SEO, Moz, etc.” But suddenly, you’re overwhelmed into inaction because you can’t see the forest because of the trees.
It reminds me of a book called Essentialism. Have you read that book?
Yes. I believe in it. It also has a life philosophy dimension to it.
Have you applied that book and its precepts in your life personally?
You have to make much more deliberate choices on the life side. Over time, I always say I have three jobs with the blog, the book, my time at Google, doing keynotes, and writing a newsletter that reaches 50,000 people. It’s very important for me to figure out how to make time for my family and be there for my children and my wife so I’m a good partner.
As my life has become more complex and busy, I’ve used the book’s philosophies to determine what is essential. I am very deliberate. I only have X number of friends and zero hours watching TV, except my San Francisco Giants – if they’re playing. I am making very deliberate choices so that I can still be able to do all the things I want to do.
I believe in this peak philosophy. There have to be things that you commit profoundly and deeply. That has also been helpful.
On the other side, in my professional job, one of my big practices is not to tell people exactly what to do but to teach people frameworks about how to think. It is far more important to help people think more sophisticatedly.
One of my big practices is not to tell people exactly what to do but to teach people frameworks about how to think.
I love building frameworks. I have these ladders of awesomeness. I’ve got the impact matrix and all these frameworks because people can figure out what is essential for them, which might differ from their direct competitor or person in a different department. My professional practice, using frameworks and how to think, has been how I contribute to other people finding their essentials.
Yeah, that’s awesome. One thing you mentioned a couple of times was being deliberate. I’m curious about your take on the term discerning as an alternative to deliberate because, ‘deliberate,’ you just go off and do your thing. You’re making decisions. You’re being thoughtful and not just winging it. But discernment has a different vibration to it. There’s more introspection, and there’s more magic and intuition implied in that term versus deliberate. I’m just curious to hear what your thoughts are on that.
I would support that. The thought you triggered with your description of the beauty in the word discernment is I just took a new role a few months ago. So I have this thing I call the Avinash user manual. It outlines how I work, my values, and how I behave. One of the things in the Avinash user manual is strong opinions loosely held.
When you mentioned discernment, I liked this idea of being curious, thinking, and changing your mind. In deliberating, you’re looking for that one thing or path, while discerning, it’s like loosely held strong opinions, which is I’ll be able to change my mind if I get new information and look at it with a more open mind. I’m open to new information.
Deliberate has a locked-in nature to it. Like, “Oh, we’re going this way, and then we’re in.” Would you change your mind if new information became available? Maybe, maybe not. I support your nudge on discerning. There’s something about strong opinions loosely held.
Something about loosely held strong opinions makes me a little uncomfortable. One of my precepts in life, the overriding principle in my user manual, is the willing suspension of disbelief. It’s contrary to the concept of strong opinions loosely held.
You should be very open-minded to realizing when is the time to time to listen to the data.
What I tell people, rather than suspending belief, is that you should have a very active imagination about what is possible. I noticed in my professional practice, and perhaps I’ve been uniquely cursed with this, that people suspend imagination. That’s not that hard. They don’t know when to give up or when to keep going.
I think you should imagine crazy possibilities and have a go at it. For example, one of my big principles is to put 15%–20% of your budget into trying new insane things because it is a way to learn new things, try new things, and be open-minded to what is on the horizon that you can’t see.
What I am picky about because of the space that I occupy in terms of being an analytics person is that you should be very open-minded to realizing this is the 30th time you have failed. At that point, it’s time to listen to the data. For some people, it might be the third time. But if you suspend belief, and for the 99th time, you’re still failing, I don’t know that you should suspend your belief all that much.
Einstein did say that insanity is doing the same thing over and over again, expecting a different outcome, right?
Yes. I suggest suspending your belief in a professional context and applying it. Try new things, explore, and take risks. Just be open-minded to new data. If you see it, change your mind.
I leave the judgment to you if you change your mind after three or sixteen times of failure because I don’t know how rich you are. You may have so much money that you can fail ninety times, and it’s no big deal. Or you have so little money that after the third failure, it’s time to unsuspend your belief because data is nudging you in a different direction.
The willing suspension of disbelief is my mantra, not the willing suspension of belief. I want to ensure that I’m not putting myself into a box and not seeing the bigger picture because I have these preconceived notions holding me back.
Refrain from being locked into the knowledge you already have and into a specific perception of consumer behavior.
Yes. In a professional context, the way I would apply it is, don’t be locked into the knowledge you already have. Don’t be locked into a particular form of doing things. Don’t be locked into a specific perception of consumer behavior.
More in the marketing context, I’m applying this. Just being open-minded, able to think differently, try different things, maybe throw crazy ideas inside your team. Use data to guide how much farther you want to go or how short you want to make that journey.
Yeah. There’s an adage in management that you make up your mind quickly and change your mind slowly. Do you subscribe to that principle?
I’m fine with variations of these things because different applications are wiser in different business contexts. For example, let’s say you and I are in a mature business. The fight is about just getting 0.01% of the market share, that’s worth a lot of money, which means you guide things a certain way. Or you and I are in a startup, and we’re fighting against big behemoths and trying to get the first 7% of the market share. I would apply very different principles to that situation.
There’s no universality to it for me that I would prescribe. It’s a really wise thing. Think about the context that you’re in. Does this make sense in the context I am in? In a lot of contexts, it will make sense. But occasionally, you might be in a context that doesn’t make sense. Let that guide the choice.
Yeah, it makes sense. What’s an example of spending 15%–20% of your budget on insane things? What are you doing that’s insane?
So many things. It varies a lot. For a client I’m working with (they’re like a B2B company), figuring out how to use TikTok feels utterly insane to them and a little bit to me. But the reality is if we want to influence, especially solve their brand problem, exist on TikTok.
We’re experimenting with a deliberately good sum of money because you don’t want to do a thousand-dollar experiment. For them, committing to ensure they have a decent enough impact is several hundred thousand dollars, or else you fail. Either way, the signal needs more money.
We’re doing something for them and trying it on TikTok. Three weeks after, the signs show we shouldn’t give up, which is good. It takes a while, of course.
On the other hand, one of the arguments I often get into comes from this realization. About six to seven years ago, I spent a lot more time on brand marketing, solving brand marketing problems versus performance marketing, which is where I grew up. That’s where I spend most of my time. In my books, maybe 80% of the content is about performance marketing.
For brand marketing, roughly 60+% of the success is determined by the creative. Not the audience, the targeting, the bidding, the frequency, not all that stuff. It’s the creatives.
My risk-taking has been trying different strategies to build creativity these last few years. No more bunnies, babies, dogs, or anything distracting from the brand. Should we do value-based advertising to get an unaided brand recall to move? Should we try celebrities? Or just different businesses and contexts?
I’m spending a lot of time taking risks and rethinking creativity because companies are very locked into their view of how they do creativity.
For the entire history of our entity, our logo has been blue. What happens if you make it black and close to invisible? I don’t know, crazy things.
I’m spending a lot of time taking risks and rethinking creativity because companies are very locked into their view of how they do creativity. The best example of that is Apple. It doesn’t matter where you see an Apple ad. If you see an Airbnb ad often, you first think it’s an Apple ad because the ethos of the brand expression is the same. But now, I’ve trained myself to differentiate the two.
I’m taking a lot of crazy things to be able to communicate the message we want to display as a brand. I’m moving hard KPIs like unaided brand awareness, one of the hardest possible in marketing. Creativity plays such an extraordinary role. It’s where I’ve spent a lot of time because if I can solve the problem, I’ve solved the contributor of 60+% of the success. It’s the one thing that has extraordinary importance. That has been a lot of fun.
It’s also an area with strong egos and opinions. Chief Management Officers (CMOs) are married to their creatives. They love their font, color, and values. They love the values and attributes.
Because strong opinions are strongly held.
Strong opinions, strongly held. Take a risk because if you win, the profit is magnificent. It’s a sustainable competitive advantage.
How do you measure the KPI of unaided brand awareness?
There are many ways to do it. On some channels, you can’t measure it. For example, if you’re doing brand advertising on YouTube, then you’re using the YouTube reports, the BLS reports that come out of YouTube. They can only measure aided brand awareness. To measure your YouTube unaided brand awareness, you must put something custom in place with an Ipsos, Kantar, or something.
Usually, it’s true test and control. We do this for TV, radio, and YouTube, and it doesn’t matter. We normally break the audience into test and control groups, and then in the test group, we’ll deliver the brand advertising. I will measure the brand lift in both cases, only the brand awareness. I will have a causal relationship back to advertising.
We’ll be able to measure three metrics at the end of the campaign. What was the percent of lifts we got, the number of individuals lifted, the cost per individual lifted, and its true test and control? Which means it is an incremental measurement.
Is it expensive to run that test or analysis?
It depends on what you think is expensive. Usually, it’s a very tiny percentage of a campaign. So the measurement is not very expensive. What is expensive is to make sure, either at an agency or in-house, you have access to a researcher who can set up this experiment properly and is analytically savvy. Most of these measurements fail because you didn’t set up the correct test and control groups.
In some cases, let’s use a KPI consideration, which you can measure on Facebook and YouTube. YouTube and Facebook are excellent at setting up your test and control. They’ll figure out the audience that is meant to get your ads. They’ll hold back the ads from a certain percentage of the audience, show them to the rest of the audience, and measure to test and control.
Using platforms like Facebook and YouTube, they take care of everything. But if you measure on some platforms where the KPI is not measurable or want to measure between Facebook and YouTube, you’ll have to set up custom measurements. In that case, having access to either an agency or an in-house person who understands how to do qualitative measures will ensure that you set it up well. But the measurement cost is very tiny in your total campaign cost.
If you measure on some platforms where the KPI is not measurable or want to measure between Facebook and YouTube, you’ll have to set up custom measurements.
Can you give an example of where the test and control groups were not properly set up?
For example, there are some channels where they’ll say that, rather than figuring out all the people who would have seen your ad and then segmented in real-time, what we’ll do is we’ll measure your current KPI before we run the ad. We’ll run your entire ad, then at the end of the campaign, measure what happened. We’ll measure before, after, and during. The problem is, there is no causality in there because a whole host of things could affect it, so it’s like setting it up badly.
For example, we’re going to look at the ages of people. People find cities where there are a lot of 20-year-olds, and then we’ll put it to the test. We’ll find another city with more 20-year-olds and make that control. That’s not a very smart way to test and control. Consider things like your current demand, your current customer base, the visitors to your website today, and where your retail presence is, and you usually use twelve or thirteen factors to create a clean test and control.
That’s why it’s not expensive from a doing perspective. But having access to a person who’s done these things in the past and paying them $500 or whatever the amount of money they asked for, a very small amount of money to set this team for you, is very good because true test and control are hard to do, especially if you don’t have a background in design of experiments.
What do you think about the typical setup of Google Analytics for a company? Is it set up wrongly? Is it set up incompletely? Is it not tracking the right things? What’s your take on most people’s implementation of Google Analytics?
Thousands of consultants now do Google Analytics implementation. Usually, if you have just implemented Google Analytics, the challenge is to ensure completeness. All the pages are tagged that you’ve delivered through GTM (Google Tag Manager) or a different company’s tag manager, so it is easy to implement and ensure completeness. But, if you’re new, completeness is the problem.
For most companies, unless they hire somebody who doesn’t know how to do it, there are also accuracy problems. But it’s less than less because most people hire consultants for very little money. They’ll implement it for you, and all that is good. The big challenge for clients I work with is ensuring that you’re extracting all the value out of the tagging process.Email signups, downloads, creating an app, and writing reviews become micro-conversions, which add value to your business. Click To Tweet
Have you set up ecommerce? Have you set up all your micro conversions? Are you firing the right set of events you need and ensuring that your campaign tracking tags, the dreaded UTM parameters, are implemented across all your campaigns? Again, it’s ensuring the things you can do on top of the basic tagging are done so that you can get 80 points out of 100 values versus 20 points out of 100 values of Google Analytics.
It all comes after the initial tagging because you can track so many incredible things you can track with Google Analytics. There are so many features that are built in. For example, data importation you can do, you can visit a loyalty now, and you can do lifetime value now. To check out the whole product page to cart, abandonment visualization in GA4 is amazing. It will completely help you change how your merchandise, but they all have to be set up after your initial tag. People leave a lot of money on the table there.
You get the first 20 points. After that, you may get to 50 easily. But after that, there’s a lot of value to be extracted. The value extraction is to understand what the business needs and getting a more advanced consultant to help you with, or your in-house individual.
By the way, this applies to Adobe Analytics, and it applies to PivotPro. It doesn’t matter what analytics tool you’re using, but just know that the standard implementation will get you 20 points out of 100. You will try and end up around 60 to 70 because then you’re cooking with some gas.
Are you going to do an update of your book, or one of your two books or both about GA4 and all the changes?
I probably won’t be on GA because so many people are far better than me at writing books about Google Analytics. My books were more about web analytics. What I used to find out there was more about how to apply different thinking no matter what tool you use, so how to use segmentation, attribution, and incrementality. The books are about making sure that you are thinking and doing analytics in a way that extracts the value for you.
Many screenshots were about Google Analytics because it was the only free tool at the time. If I used that screenshot, anybody could replicate what I was trying to do. They could build the custom reports I built, but they were not about Google Analytics. My friend Brian has written fantastic books about Google Analytics. You should buy all his books. They’re way better than mine on GA.
One of the reasons I haven’t written it in the book—the second book is ten years old, I think—is that I started writing the newsletter, and then it has 55,000 subscribers. After the George Floyd protests, I decided I wanted to do something. So what I did was I made my newsletter. It’s like everybody pays me to write this letter.
I decided to donate all the money to charity, 100% of gross revenue. The first six months were about $70,000–$80,000. Last year was the second full year, it was $200,000 gross revenue, and I donated it all to charity.
It turns out that I can make so much more money using the newsletter because books don’t play, well that I decided to write the newsletter once a week and let people pay for that. But since then, I’ve been repeatedly told that my books reach more spaces than the newsletter ever will, so I may write a book. I just haven’t had time.
There’s an expression I learned from something called nonviolent communication. The expression is “Don’t should all over yourself.” If you feel called to do it, if you feel like it’s going to spark joy for you to do an update to the book, then do it. Otherwise, it’s a chapter in your life that’s now finished.
Over the last few years, my set of interests has branched into more sophisticated marketing and marketing strategy. I’ve written a lot in my newsletters about that, and then course, core strategic analytics. I have enough content. I should take a break from my work for a couple of months, take all the content I have, find all the missing pieces, and put it into a book because it does reach more places than my newsletter ever would.
That’s great that you’ve donated all your proceeds from the newsletter to charity. Have you done a back calculation to figure out, like if you want to donate a million dollars, that means you need these many subscribers, this price for my newsletter, and so forth? I’m guessing you haven’t raised the price of the newsletter since you unveiled this. Maybe it’s time.
I wish I were that smart. In both books, I’ve donated all gross revenue to me in both books, which is 17% of the net from Wiley, to charity. Between the books and the newsletter, about a million dollars were donated to charity.YouTube and Facebook are excellent at setting up your marketing test and control points. Based on the results, the sites’ algorithms target your ideal audience. Click To Tweet
When I started the newsletter 4½ years ago, and the paid one has been 2½, I just said, “Oh, everybody charges $10 a month and then $100 a year.” I was like, “Okay, that’s it.” I haven’t thought about it since then.
The only time I have to think about it is when I have several companies that subscribe. They’ll say, “Okay, do you have a group subscription? If you have advice, I’m happy to take it. I do it randomly.” I’m like, “How many people do you have?” They’ll be like, “We have 70 people.” I’m like, “Okay, 30% discount.” I’m making it up, but I think I should spend a little more time thinking about how I should do group discounting because I do it randomly right now.
Look at the data, or at least your intuition.
I’m very honest with people when they ask for a discount. I said, “Look, all the money is going to charity, 100% of gross revenue. I will cover all the costs myself. Take whatever discount you want. It’s more money for charity anyway.”
That’s cool. What’s your take on Google Analytics 4?
I’m very excited about it. I have to send it out. Several years ago, I spent time with the GA team about what GA4 should be, and I haven’t been involved since. The team is amazing. They’ve built an amazing product.
I broadly think that it is a much better update. It fixes a lot of challenges that I think were not so great. For example, one of the big shifts between GA3 and GA4—I don’t know that we ever called it GA3. Anyway, the old GA and GA4—is that GA4 is far more demanding of analysis than reporting, so less data puking and more asking you to go in and do analysis, which I’m very fond of.
There’s an entire section for you to do that analysis because every business is so different that I’ve been a big fan of asking your questions versus getting a bunch of answers and figuring out what questions they answer for years, which is the old model. I’m very fond of that.
The visualizations are amazing. The churn rates, lifetime value, and integration to BigQuery, they’re all amazing. I know this has been controversial for many people. Still, I am very fond of getting rid of all the various attribution things because it was never proven that even distribution, first click, or any of these were good.
You should be on DDA (Data-Driven Attribution). It’s a machine learning–based model. It is built custom for every site based on your customers and behavior. That’s the way to go. I’m glad that’s the choice because many people used all six or seven models without thinking about it.
I just read somebody who’s like, “Our business runs in first-click attribution.” I’m like, “Do you realize that first-click attribution is like giving your first girlfriend all the credit for you marrying your wife?” It’s really that silly. But that’s what they’re doing because it is there. They can use it. They somehow justify that to their business, but that’s such a silly thing to do.
I’m glad that they’re doing that. Not everything is working. There are still teams that are going to get out. It’ll get there. But as a fan of somebody who advocates for analysis, I’m excited about what’s in GA4.
What about Google Tag Manager? You mentioned earlier that having Google Tag Manager tied into their implementation of Google Analytics is a good thing. What about those companies that aren’t using GTM and only adding their GA tags just by pasting it into, the Yoast SEO plugin for WordPress?
In a way, the Yoast plugin also does what a GTM does. It just takes care of making sure that site is tagged. Using something that is a tag manager–type solution, whether it’s using Google Tag Manager, Yoast, or some other plugin or company, is a better way to tag because you only have to make changes in one place, and the rest of your site automatically gets tagged. It’s a much smarter way to do it versus doing it manually.
The only thing I recommend not doing is manually and individually tagging all pages because it becomes a nightmare for you to update because you are out of sync very quickly. Using any tag manager is a better way to do this because it will ensure that if you want to make changes, you want to make updates, you want to improve things, that is completely possible in a very simple way.
You briefly mentioned micro conversions. What would be some examples of micro conversions for business-to-business and business-to-consumer?
Even if you’re an ecommerce website, your conversion is probably 2%–3%. If it’s 3%, it’s probably way higher than any other site. So the number of people converting to your site will be very small, but the rest is a wasted opportunity.
The three areas that are great micro-conversions are Google, Bing, and Facebook. Have a very healthy way of figuring out what to do with them, especially Facebook as the most important.
When I was writing my first book, the encouragement I gave people was that the rest of it was not a wasted opportunity. We should figure out if there are other things of value that people are doing on our website if they’re not converting.
Email signups, downloads, going and getting our app, writing reviews, all of these things become micro-conversions, which is they’re adding some value to our business. In some cases, they will help somebody convert later in the future. Or things like reviews, which are very good because they increase the product’s conversion rate overall. These are some examples of micro conversions.
You can say if I put the macro conversion at 2% and the micro conversion, then basically, for 20%–30% of the people who come to my site, there is something of value happening. That’s a very different way to think about the full possibilities of what is on your website versus saying 2% converted and 98% wasted. That’s not the reality.
If I’m a B2B company, the macro conversion often for B2B companies is that you submit a lead. You came to our website, you submitted a lead, and that’s great. Now go into our CRM (Customer Relationship Management) system, our salespeople will call you, and so forth, and you will get converted.
If I’m a B2B company, it’s not unusual to say, I also have written a lot of white papers, I’ve done videos and webinars about my products, I’m asking you to sign up for my email newsletter, and I have the latest promotions on my website if you do X or Y. So some of these things would end up causing micro outcomes for a B2B business.
In fact, for one of the B2B companies I was working with recently, their big macro conversion was not the fact that you submitted a lead that was a micro conversion. Instead, they wanted you to come to their in-person events. Usually, I do event signups as a micro conversion. But in their case, it became a macro conversion.
Things can flip, but the full collection of things that define success for your website, identifying them, making sure they’re tagged, and they can track them, allows you to be a much more sophisticated digital marketers because we will figure out, oh, for this micro conversions, Google is great. Let’s spend time doing SEO and PPC, etc.
For these three great micro conversions, it’s Bing. It’s Facebook that is far more important. So now you have a very healthy way of figuring out what you should do with Google, Facebook, and Bing.
It is incredible what you can do once you track the micro-conversions of users’ reviews. You can figure out a way to engage your audiences, which can greatly impact improving conversion rates.
I’m using a very simple example, but it allows you to have a more informed and robust marketing strategy versus saying, I have a 2% macro conversion rate. If somebody’s not converted, everything else is a waste of time. It just makes you a more sophisticated and savvy advertiser.
I remember Mike Moran explaining this when he was working at IBM. He would measure white paper downloads. They did a study to determine that each white paper download was worth $100 based on an 18-month journey until the big consulting gig closed and the new client came on board.
If that was a seven- or eight-figure contract, and then they back-calculated what the white paper download was worth, let’s say it’s $100 just for round numbers’ sake, that would be an example of a micro conversion. That’s a lot easier to track than an 18-month buyer journey. I’m curious to hear of any case examples that came to mind when I shared that one from Mike. Anything else that might pop into your mind?
Mike’s fantastic. I’ve had engagements with him, and he’s amazing. There are two concepts. One is tracking the micro conversion and setting the goal value for the micro conversion. If you do that, you can unlock attribution and many other reports inside tools like Google Analytics or other tools in the marketplace.
How do you figure out the goal value? What’s the worth of an email signup? We can follow Mike’s process, except it is shorter for ecommerce and other entities and longer, but you apply the same process. You have everybody’s email. This way, you have at least the email address, and then the CRM system, you have the email, then you can track the conversion at some point, somebody faster, somebody slower.
For me, the most fun was taking a relatively large ecommerce website and figuring out the value of reviews. This was a few years ago, so forgive me. This is not precise, but after the seventh review of a product, there was a statistically significant increase product’s conversion rate of the product if it had seven reviews or more. The most surprising thing was it didn’t matter if the reviews were positive or negative. They were all crap. Of course, it impacted it.
Usually, that rarely happens. You always end up at four points or something. It doesn’t matter. People just don’t like something, and many like it. So throw that little outlier out unless the product is just complete garbage. After seven reviews, we found a statistically significant increase in the website’s conversion rate.
We were able to use that incremental lift and compute the value to the business of every review. That was super cool. That mindset has been implemented on all kinds of websites.DDA (Data-Driven Attribution) is a tailored model intricately crafted for websites that leverages customer data and behavior to shape their insights. Click To Tweet
My favorite is Rent the Runway. Their reviews are the best. Of course, I haven’t seen the entire internet, so I can’t say. But from all the websites that I have seen, I love how Rent the Runway uses reviews because their users are so passionately engaged. They post photos, and they post videos.
It’s really fun that I’m trying to rent a dress for myself, and I can see 7–8 women who have ordered the dress, who are all different sizes and colors. I can see how it looks on them. It helps me figure out if I should order this dress because I can’t see that much variety and diversity in the standard product shots for the dress.
It is incredible what you can do once you track the micro conversions of reviews and then figure out a way to engage your audiences. It can have a huge impact on improving conversion rates. That was one of my fun ones.
Yeah, I like it. I wasn’t even aware of Rent the Runway.
It’s for women’s clothing, although they now have a men’s brand. But I’m very fond of the original Rent the Runway for women. They do reviews exceptionally well.
That’s cool. Earlier, you mentioned frameworks, and you have your awesomeness framework or something like that. You have your impact matrix and so forth. Do you have any frameworks that our listeners would benefit from regarding AI?
It’s a very disturbing, well-done, dark movie.
It’s like everything is AI. Paintings are AI. Sofa design is AI. Every agency is like AI reports. I’m a little tired of that. So in the context of businesses, I’ve created these two frameworks. The first framework is AI as what.
If you begin any discussion of AI, it’s really important to understand what we will use AI for. The framework simply outlines AI as a tool, AI as a co-pilot, or AI as a means. At the moment, I’ve put all use cases in a business context into those three categories.
The newsletter outlines, first, how to figure out which of those three use cases apply to you. It’s very important. Then the second framework is the AI activation framework. So what I’ve taken is, from a business context only, I’ve taken all possibilities that I humbly can see at the moment, put them into five categories, five clusters, and then I rank ordered the current in the order of profit you can generate from AI in five clusters. So they start with one, two, three, four, five, and I say, do this first, then do this, then do this, then do this, then do this.
As you go from one to five, generative AI is one, and NLP (Natural Langues Processing) is five. It means that by using generative AI, you can create profit for your company today by delivering better experiences, etc. With NLP, it’s the most out, where a lot more work needs to happen before everybody can monetize the value they get from it.
That’s awesome. That sounds good.SEO is always evolving, and it’s vital for your business. It also stands as the foundation for tailor-made AI answers and future possibilities. Click To Tweet
Those two frameworks might be helpful when you think about AI so that you’re not wasting time because I see lots of companies just get on the hype train, run, and spend a lot of money, then, in In the end, they have less money and no AI.
AI is certainly making a lot of people nervous. They’re thinking, oh, with SEO changing, should I even invest in SEO? Is SEO dead now that generative AI will be part of the search results? Are the ten blue links going to go away? There’s a lot of fear out there. What is your position on this?
There is a need for SEO. The way we practice it is evolving. For example, one dimension of SEO is understanding what we stand for, our purpose, focus, and the content categories we should be creating.
In technical SEO, we’re like, “Is the site loading fast? Is the URL good?” All the technical bits and bytes, all that remain relevant. What is happening that is more productive is SEO now having an initial helping hand with AI possibilities.
For example, one of the things we want to do with the series is write really good content. If there are ten people selling car insurance, you all go to AI and say, write me articles on car insurance and all the crap, or there at least be saved. Sometimes also be crap on top of being saved. If you’re an SEO, you can say, I want to write about car insurance. You can get initial article ideas from it. It’s AI as a muse.
You can say, “Write me a 300-word article on a podcast I’m going to publish about analytics,” and it will write you a 300-word article. But you still need the human, the SEO person, or someone else at the company to read it and say, is this true? Does this make sense? Is this right? Oh, the last two paragraphs are bad. I’m going to rewrite it. But thank you very much for making my life easier by one hour. At the moment, emerging SEO possibilities are that it will find those things.
We are using AI to collect lots of data to find technical anomalies in how we deliver the user experience to see if we can improve the technical SEO by finding patterns to fix at a very large scale.
The second thing that I’m also very excited about is the technical SEO data our company is collecting, Croud, the agency I work with. We are using AI to collect lots of data to find technical anomalies in the way that we deliver the user experience to see if we can improve the technical SEO by finding patterns to fix at a very large scale. Working with a very large client, there are enormous amounts of data. Things are not quite as cut and dry, and there are patterns that we want to monetize. It would take too long for humans, even if they could figure it out.
We’re building this tool called SEO-Max. You don’t have to use various algorithms to allow us to do technical SEO better. It’s another place where the machine learning algorithms we’re building are helpful to the SEO person in the practice of search optimization. It remains very important.
There is uncertainty if we’re heading to a space where we go and ask Bard or ChatGPT and Bing to plan a vacation to Croatia, and he’s just going to do the whole thing for me. “Where are my link packs? Why are people coming to my blog about plans to Croatia? And how is all that going to play out?” I’m afraid even if you ask Google and Bing, and they don’t know what the hell’s going on or where things will end up. We are certainly in a period of uncertainty.
I just met Microsoft, their team working on the new experiences. I asked them, how are you going to show ads? If you wrote this tight article about vacations to Croatia, whose ad do you pick? Are you going to have three ads like Google and Bing do? On there, they said, we’re doing a bunch of experiments. We’re not sure how the SEO links will come out.
Yesterday, I was playing with one of the experiments on Bing. They put these parentheses at the end of almost every other sentence. It’s linking back to the source. That could be the new SEO link strategy. But it’s so new that I am still determining from my humble vantage point that even the search engines know exactly how it will play out, but the next six months will be very interesting.
The SEO parts will play out well because they’ll figure out this new Bing experiment and how to link back to the sites more frequently. But how do the ads show up? Should three ads show up if the idea is I can give you the one perfect answer? I’m not sure how this is going to play out. There is even more uncertainty on the paid site.
Interstitial. Please wait. For ten seconds, watch this commercial, and now we will answer your question.
It reminds me of what’s the Wi-Fi provider I hate that makes you see the ad first before they give you Wi-Fi. New York has them.
Who knows what’s coming? Anyway, let’s hope for the best.
Yes, exactly at the moment. But remember, the art and science of SEO remain important. SEO professionals must get very familiar with these emerging technologies because it makes their job easier. We’ll figure out what else to do as the search engines evolve. I don’t think SEO is dead. I think SEO professionals have had very profitable careers ahead of them.
I would agree with that. Thank goodness, we have a fourth edition of The Art of SEO about to be called.
New updated with more AI.
For sure. This was such a joy. I’m so glad we got to reconnect. It’s probably been four or five years since we saw each other. I think it was at the Semrush event.
Anyway, thank you so much for taking the time to share your brilliance, wisdom, and experience. The newsletter is a great starting point if our listener wants to learn more from you. What’s the best URL for folks to go to to learn more from you?
Kaushik.net, or just type Avinash into Google. It’s not that hard to find me.
You’re the number one Avinash on the Internet.
Sometimes, yes. There’s an Indian actor far more famous than me, but it’s not hard to find me if you just type in Avinash into Bing or Google.
Awesome. Thank you so much, Avinash. Thank you, listener. Have a great rest of your week. We’ll catch you in the next episode. I’m your host, Stephan Spencer, signing off.
Facebook – Avinash Kaushik
LinkedIn – Avinash Kaushik
Twitter – Avinash Kaushik
Your Checklist of Actions to Take
Embrace a data-driven mindset and use analytics to guide my decision-making process. This will help me to assess how to utilize actionable data.
Spend more time on collecting quality rather than quantity data. Focus on the data that is essential to my business.
Utilize data to gain insights into user behavior and optimize your website’s user experience accordingly.
Invest in ongoing training and education to enhance my data analytics skills and keep up with industry trends. Familiarize myself with emerging technologies – this will make my job easier.
Focus on delivering unique, valuable content that satisfies user searches and addresses their needs.
Use data visualization techniques to effectively communicate insights and make data more accessible.
Pay attention to the technical aspects of SEO, such as site speed and indexability, to improve search engine visibility.
Monitor and analyze key performance indicators (KPIs) to measure the effectiveness of my marketing efforts.
Stay up to date with the latest SEO trends and algorithm updates to adapt my strategies and stay ahead of the competition.
Gain valuable insights about data solutions, digital experience, digital performance, and more by visiting Avinash Kaushik’s website.
About Avinash Kaushik
Avinash is the global Chief Strategy Officer of Croud, a leading full-service marketing Agency.
His prior professional experience includes a sixteen-year stint at Google, and roles at Intuit, DirecTV, Silicon Graphics in the US & DHL in Saudi Arabia.
Through his newsletter, The Marketing < > Analytics Intersect, blog, Occam’s Razor, and his best-selling books, Web Analytics: An Hour A Day and Web Analytics 2.0, Avinash has become recognized as an authoritative voice on how marketers, executives, teams and industry leaders can leverage data to fundamentally reinvent their digital existence.
Among the awards Avinash has received are Statistical Advocate of the Year from the American Statistical Association, Most Influential Industry Contributor from the Web Analytics Association, and Founder’s Award from Google.