Select Page

In this episode of In-Ear Insights, Katie and Chris dig into the announcements around Google’s newly announced AI model for advanced search, MUM, or the multitask unified model. Learn what MUM is, why it should matter to content producers, and how to think about adapting your content strategy to deal with what could be a very big change to SEO.

Subscribe To This Show!

If you're not already subscribed to In-Ear Insights, get set up now!

Advertisement: Data Science 101 for Marketers

Do you want to understand data science better as a marketer? Would you like to learn whether it’s the right choice for your career? Do you need to know how to manage data science employees and vendors? Take the Data Science 101 workshop from Trust Insights.

In this 90-minute on-demand workshop, learn what data science is, why it matters to marketers, and how to embark on your marketing data science journey. You’ll learn:

  • How to build a KPI map
  • How to analyze and explore Google Analytics data
  • How to construct a valid hypothesis
  • Basics of centrality, distribution, regression, and clustering
  • Essential soft skills
  • How to hire data science professionals or agencies

The course comes with the video, audio recording, PDF of the slides, automated transcript, example KPI map, and sample workbook with data.

Get your copy by clicking here or visiting TrustInsights.ai/datascience101

Sponsor This Show!

Are you struggling to reach the right audiences? Trust Insights offers sponsorships in our newsletters, podcasts, and media properties to help your brand be seen and heard by the right people. Our media properties reach almost 100,000 people every week, from the In Ear Insights podcast to the Almost Timely and In the Headlights newsletters. Reach out to us today to learn more.

Watch the video here:

Can’t see anything? Watch it on YouTube here.

Listen to the audio here:

Download the MP3 audio here.

Machine-Generated Transcript

What follows is an AI-generated transcript. The transcript may contain errors and is not a substitute for listening to the episode.

Christopher Penn 0:02

This is In-Ear Insights, the Trust Insights podcast.

In this week’s In-Ear Insights, the times they are changing recently at the Google marketing live and on Google’s blog itself, they made it a pretty big announcement saying that they’re going to make search organic search and search performance 1000 times faster.

And the way they’re going to do that is with something that is called mum.

It’s a type a new type of machine learning model that is multiple formats, and uses all Google’s experience in natural language processing and document matrices and stuff to return.

more intelligent results.

The goal that they’ve said in the blog post and academic papers, is for Google, to help a dilettante behave like an expert.

So the example I gave is, if you were typing in, you know, what does it take to climb Mount Fuji? It would pull in all this data.

And then the next question asked is, what would it take to climb Mount Kilimanjaro and using the understanding of context, it would help you understand the differences between climbing these two different mountains, types of gear you would need and things like that.

So Google is effectively trying to create this multi domain, multi format model of search, and the technical underpinnings, they’ve published in a couple of papers that the folks over at Search Engine Journal were able to dig up.

And it’s really fascinating.

I don’t know how, how bloody we want to get into the tacticals.

Katie, but when you hear 1000 times faster and 1000 times better.

And knowing that it’s Google, they can probably back it up.

What is your What are your first inclinations as a marketing manager, as a, as a marketing executive, say like, hey, looks like SEO is gonna be changing a whole lot.

Katie Robbert 1:51

My first inclination as you’re describing it, is, well, that content better exists that Google’s looking for.

You know, and so I think that that is where as a marketer, I’m less concerned about sort of like the going deep down the rabbit hole of exactly how it works.

And just sort of knowing the broad strokes of what does this mean, for me, as a marketer? What do I need to be concerned about? And it sounds like, and correct me if I’m wrong, it sounds like the thing that I need to be concerned about is tightening up my SEO strategy, and making sure that I have a variety of content around those specific things that people are trying to find me for answering those questions of what where, how, when, why.

So that, as this new and improve search engine is finding answers to questions.

I’m the one being found if through if Google is basically taking a 360 approach to answering the question, I want to make sure that I’ve covered all 360 angles on one particular piece.

111 topic? Is that right?

Christopher Penn 3:02

That’s a really good way of thinking about it.

Well, I’m gonna share my screen here.

For those who are listening on audio, you can go over to the Trust Insights, YouTube channel TrustInsights.ai dot AI slash YouTube to see what is this is from one of the academic papers that Google has published on this topic.

And it’s a good quick summary of how things change on the left is someone asking a question and Google going through it’s it’s four or five stages of information retrieval today goes to its index retrieves information, it ranks it, and it spits out results.

And the new version that they’re proposing.

There’s an existing gigantic, pre trained model.

And your question goes, just go straight to the model, it spits out results with cuts down on many of the steps that would previously go into what they call the older version of search.

And without getting too deeply into transformer models.

Fundamentally, what they’re saying is, they’re going to have search be the model itself, as opposed to your models operating on a database of search results.

And so Katie, your point about that content had better exists, it’d better be high quality, it’d better be rich, heavy, it’d be multi format, or all things that would be taken into account into this kind of newer model.

The best way to think about it is it’s like the difference between, like, going to local hardware stores versus going to Amazon, where Amazon just has everything, it may not necessarily be the best price or where you know, the most convenient, but it’s got everything could possibly, you know, look for if you’re looking for a half inch washer, you know, flash washer, you can get it there and your local hardware store baling habit.

The older version of this search model that Google’s talking about is essentially like all these different stages of it are kind of like those individual hardware stores where the new versions like yeah, we just put everything in one big warehouse.

And because AI is powering it, it can get you stuff faster than you know obviously going to a real world warehouse.

But from an SEO perspective, it means that we’re gonna have to radically rethink about how we build content, we’ve got to radically rethink about how our content is contextually related to other things.

Because in a model like this in this in this model format, there is there’s no, the crawl stage still has to happen, something, Google still has to go out and crawl your website and pull all the data from it.

But instead of, you know, indexing and retrieving and ranking, it’s just going to build a mettam, it’s just going to go into this massive model.

And anytime somebody searches for, say, analytics consulting, it’s going to take into account obviously, the things like human quality, it’s going to take into things account things like you know, links to some degree, but document type and contextual relevance is going to be a lot more important, which means that our quality of content, and what we publish has to get a lot better has to be a lot more authoritative.

It has to be longer probably, and has to be in multiple formats.

Katie Robbert 6:01

So it sounds like the first place to start I mean, so again, as a marketer, I’m thinking, so what, what, what, how does this directly impact me? What do I need to do first, so the first thing I would do based on what you’re describing, is make sure that my I would do my technical SEO audit, I would make sure my website is function, I would make sure Google can find it, I would make sure that my sitemap is up to date, I would make sure that I don’t have broken pages and broken links and all of those things that Google would ding me for in terms of, you know, what is technically on my website.

So I would do that first.

And then second, I would start to think about the actual content that I’m producing.

And so this is something that you know, that we’ve talked about, you know, a lot, which is that transmedia framework, how do you make one piece of content work for you in multiple different ways.

And you know, as a friendly reminder to people, YouTube is a Google property.

And so don’t forget to include things on YouTube.

Because that will, I would assume be taken into account in this model, since YouTube is part of Google.

And so, you know, one of the, quote unquote, easier ways to make one piece of content work for you multiple ways is essentially what we’re doing right now is to, we’re recording a podcast, but we’re also recording the video of the podcast recording, so the video will go up on YouTube, the audio will go in our podcast player, but also on our website, then we will have the AI transcript, which essentially, which is essentially a blog post, which will then clean up and sort of make it easier for Google to read.

And that right, there is at least three different pieces of content out of one recording that we’re just sitting here doing on a Monday morning talking about how Google is going to be changing how you’re searching.

And so that’s what I would do is first I would look at my website, make sure that technically, it’s all cleaned up and tightened up.

And then second, I would focus on the content itself, because if Google can’t find your website, it doesn’t matter what kind of content you got.

Christopher Penn 8:10

Exactly right.

Yeah, it’s you still need to have the crawling part be satisfied.

One thing that that strikes me knowing how long these machine learning models workers, I think, we’ve in SEO, we’ve talked about, you know, internal linking structures, and how they’ve been devalued over the years as a way to rank but not necessarily as a way to be indexed.

And I think, in terms of this new model, having a good strong internal linking strategy will help not only Google find pages on your website, but also understand the relationship of different documents to each other, like this document is linked to this document, this document is linked to this document and creating that very strong mesh that helps a machine learning models say, Okay, these are all related, like everything on the Trust Insights website is related.

And oh, by the way, you know, we’ve, we’ve set up a cross links to the marketing over coffee podcast, and my personal blog to your personal website, lots of cross linking, for your different properties will help that model go, Oh, these are all related.

These are all things that that are in some way bound to each other, even less so semantically.

From a logical structure, they’re bound, they’re bound together.

I also think there’s a very strong case to be made.

Now.

As we start thinking about a model based version of search, that you need to be doing more collaboration, more guest posts on other sites, more other people posting on your site and creating those links, again, to help a model understand what things are related to each other.

Especially when you’re talking about semantics, right? So if I go into a guest post, say on Social Media Examiner, it’s about, I don’t know Tiktok, right.

That’s not necessarily as relevant to what I want to be known for, as hosts to be doing a guest post over With links back to our site about analytics.

And so as we think about what we create content around this, it goes back to something John Mueller was saying on one of the Google Webmaster chats, he’s like, the strategy of creating clickbait content like, you know, 44 ways to, you know, 44, you know, sites to see for horror film addicts, topically, would make no sense for our website, even it would do well and get links and get clicks, it would give the model basically the wrong context for what we want to be known for.

So we also have to be very careful as marketers to be thinking, what is it that we want to be known for? And I’d like your perspective on this key to know, what do we want to be known for now? But where do we want to be going? What do we want to be known for the future? And how do you create a content plan that will train a machine learning model to make that that transition seamlessly.

Katie Robbert 10:54

So I’m going to need a couple of things.

One, I’m going to need a big whiteboard to I’m going to need markers that aren’t dried out.

And three, I’m going to need a bunch of post it notes.

Because this to me sounds like a really good planning session.

Because what you want to do is you basically want to start to break down your business into those, you want to start with those major buckets.

And so Chris, you’ve mentioned a couple of things.

So you’ve mentioned Tiktok, or social media, let’s call it social media is the big bucket, you’ve mentioned analytics, you know, we also do, you know, other types of consultants.

So I would start there.

So the first thing I would do is sort of like, on my dry erase board, I would draw out those big buckets.

And then I would start working my way down more granularly into those subtopics of like, okay, within analytics, what do we want to be known for? And sort of start at that moving it that way? And then what I would do is I would be like, Okay, what content Do we have currently, that maps to these things? And then where are the gaps? What are we missing? And so if we want to be known specifically for, you know, social media, tick tock, tick tock engagement, I can tell you right now, we don’t have any content on our website that covers that topic.

So okay, that one needs to get assigned out and someone needs to cover that topic and write about it.

And so this, obviously, like, I’m talking about a very manual process, I’m sure you would have some sort of machine learning way to do this.

But thinking about it strategically from the What do we have, what do we need, and then longer term, longer term, we want to be known more for organized organizational behavior and process development and change management, we don’t have a lot of content that cover those covers those things right now.

And so right now, it wouldn’t be the priority.

But there needs to be a plan to cover those things within probably the next six months, if not sooner.

And so that’s where my brain immediately goes, as you’re talking about that is, let’s start with, what do we have? What do we need? And then what’s that extraneous stuff that we can just sort of like, let fall by the side.

And so, you know, to your point about, you know, let’s say you did a guest, a guest podcast where you cover Tiktok, I’d be like, that’s great that you know about that.

But that’s not business that we ever want to bring in.

So that can just fall to the, to the side of like, here’s all the extra stuff.

And so that’s just going to sit there, I’m not going to focus on optimizing that.

And, you know, re sharing that out on social media, I’m going to focus on the content that we’ve created around Google Analytics for the content that we’ve created around, you know, process operations, those kinds of things.

Christopher Penn 13:38

Exactly.

I think that’s a key point that a lot of people, including myself and get uncomfortable with is is you some of your content, you got to prune.

Some of you’re going to say, yeah, this isn’t helpful anymore.

Even if it’s driving a lot of traffic.

Like this is one blog post on my website.

My personal favorite, I know I have to get rid of I just haven’t gotten around to it.

It’s the difference between two different types of nasal nasal decongestants.

I wrote it like 2007, and still like one of the most popular search posts, but you know what, that’s not my crowd.

And that’s not something I want to be known for.

So that’s one of those things that even though it drives a lot of traffic, I’m gonna have to take a short term hit in traffic, it’s gonna have to go away because with this new model, I don’t want that to be something Google Translate and says, hey, you’re getting a lot of traffic and search results for this.

Maybe you’re authoritative about this, like, oh,

Katie Robbert 14:27

let’s see.

That’s why you need someone like me whose favorite pastime is throwing things away? So therefore, oh, but I mean, in all seriousness, you do need to have someone who is not that they’re emotionally disconnected from the content but basically, they can sort of see very clearly Okay, this can stay this can go this can stay this can go sort of like going, you know, sorting through all of your stuff.

of like, okay, I want to keep these five shirts but these 20 over here.

I haven’t worn in five years.

Why am I holding on to them for sentimental value? Am I ever gonna wear them again, probably not, they’ve been out of style since the 80s.

And I would just look foolish.

Or I can just keep these five that I literally wear all the time.

And our you know, working for me is very much the same idea.

And so if, you know, I sort of, I’m going to go into sort of the human psychology piece of it for a second, if you’re someone who has a hard time, you know, organizing, you know, your bureaus or your closets, you know, getting rid of those kinds of things, it might also be difficult to do the same thing with your website content, because it’s stuff you wrote, it’s stuff you own.

And so you may have that attachment to it, that you might have a hard time getting rid of content about your sinesses, even though that is has nothing to do with what brings you money.

And so you need to start to be thinking of it in those terms.

And if you can’t, you need to bring someone along with you on that journey, who can make those decisions, but you need to relinquish control and give that person the authority to do that kind of work.

Because the more you fight against it, the more difficult it’s going to be.

So that’s sort of just like a little bit of that human psychology piece of that is there’s a reason people don’t get rid of stuff.

Like there’s a whole show called hoarders.

And believe it or not, you can’t be a hoarder on your own website, because, well, it brought in five people to my website last month, therefore, I must still need it.

You don’t, you need to be okay, with letting those things go.

Knowing that you’ll be creating more contextually relevant things that will bring in the right people.

Christopher Penn 16:35

Exactly.

Or if you don’t like throwing things away, the logical thing would be to create a separate blog, that is just all your random trash, and just move all your random trash over there and see what happens, I’ve actually got to try that we’re gonna see if I can move all my irrelevant stuff to there.

But I’m gonna put redirects from the site where it lives now to the random dress site and just see what happens.

I think it’d be fascinating.

At the very least I quit adsense ads on it.

Katie Robbert 17:01

I feel like at least as long as you’re taking baby steps, people can’t immediately take that leap to throwing things out, you need to just like, put it in a bag and move it in the hall.

And that’s at least a baby step.

So you can go ahead and do that.

God, I lost my train of thought.

So basically, the point the point being of, you know, what is relevant content? So I think one of the things one of the questions marketers probably have is, well, so let’s say I start with the topic of analytics, you know, let’s start here.

How do I then find all the spokes that go out from analytics to make sure I’m creating the right kind of contextually relevant content that’s related to that key topic?

Christopher Penn 17:51

Yeah, that’s what you need a good topic model, both for your own website.

Like I would actually say, you know, to your question earlier about, is there an automated way to do this? Yes, there is you ingest all the content on your site.

And you use a type of machine learning called topic modeling, you say, Tell me, the different topic clusters that are on my website, what do I blog about, you could do it with just like blog post titles, for example, or the whole content, the whole content would be better because Google’s gonna be using the whole content.

And just make that that structured map of here’s all the things that are, you know, that the machine has identified as as being subjects that appear prominent, and then go out and go, Hmm, that’s not what we want to be known for.

And then what you do is, because I do this all the time, I wrote a piece of software to do it.

Take the top 10 or 20, search results for that term, feed all those documents your child to your topic model and say, okay, where’s the gap? Like, I’m talking about analytics.

And I’m talking about, you know, Google Analytics and stuff like that.

And the model returns back.

You know, they talk about analytics and Tag Manager.

But when you look at the competitors, I don’t talk about, say, Google Search Console, like, Ah, that’s a gap.

Like, I can see the difference between my model and the competitors models and go, this is kind of an issue that I completely missed the boat on these things.

Now, if you do miss the boat, obviously, you want to create that content as quickly as possible.

But it also might be one of those things where you might need to do some professional development if you don’t actually know what those things are.

Katie Robbert 19:20

I think the other thing, that’s an interesting, you know, point in terms of like weeding out the stuff that is irrelevant.

One of the biggest challenges is making sure that you’re okay, not being everything to everyone.

It’s really a good exercise in focus.

And so, you know, maybe, maybe one of the relevant topics is, you know, around just as an example, like teaching Google Analytics, let’s say that’s not an area or an arena that we want to be in.

I mean, it is but in this example, let’s say it’s not, you need to be okay with making that decision to go I Don’t need to cover that angle of it because it’s not something I ever want to or need to be known for.

And this is where sort of that longer term planning is going to be helpful.

Maybe you don’t need to be known for it today.

But maybe it’s something you want to bookmark for, you know, your annual planning, to revisit to say, do we want to be known for this now.

And so I think it’s also sort of making sure that you’re prioritizing and focusing in to say, what matters today, what matters six months from now, what matters a year from now, so on so forth, and then working your way out that way, because you can’t do it all at once, and you can’t be everything to everyone.

It’s just, it’s not sustainable.

And it’s not, it’s not working smarter.

Christopher Penn 20:42

I will throw in a quick plug for last week’s.

So what marketing analytics insights live show, we talked about identifying trends and doing trend forecasting and stuff.

So if there’s some things that you’re thinking about, go go and watch that show, it’s up on the Trust Insights YouTube channel.

So just to sort of summarize, and I think we’re probably gonna spend some more time down the road on this particular machine learning model, as we start getting more details from from Google about it.

But the things you can do to prepare today, number one, you got to audit your content, figure out what you got, because you can’t make decisions.

If you don’t know what you got.

figured out the gaps, places where you should be that you’re not figured out the pruning, the things that places you are that you shouldn’t be at, but most important before any of that.

Write out a plan to figure out what your goals are today and what your goals are in the future.

Because if you don’t have those written down, all of this kind of doesn’t matter.

Anything I forgot Katie.

Katie Robbert 21:40

No, I mean, I think that’s about it.

Well, actually, the other piece is make sure that your website your technical backend of your website is working properly, and that you have a sitemap that Google can reach and can read and can index.

Christopher Penn 21:54

Exactly and spend some time if you if you didn’t catch it in last week’s Trust Insights newsletter, it’s up on our website about an introduction to Google Search Console.

It’s a YouTube video we put up that walks through the basics, so that you understand how to get started with your tactical SEO audit is up on on TrustInsights.ai dot AI slash YouTube.

If you have questions or comments about anything we’ve talked about in today’s show, hop on over to our free slack group Trust insights.ai slash analytics for marketers where you earn over 1800 other marketers can ask questions and give advice to each other.

And wherever it is you’re tuning in today.

There’s a challenge prefer to get it from us.

Go to Trust insights.ai slash ti podcast where you can see all the different ways to tune into the show.

Thanks for watching and we’ll talk to you soon.

Take care need help making your marketing platforms processes and people work smarter.

Visit Trust insights.ai today and learn how we can help you deliver more impact


Need help with your marketing data and analytics?

You might also enjoy:

Get unique data, analysis, and perspectives on analytics, insights, machine learning, marketing, and AI in the weekly Trust Insights newsletter, Data in the Headlights. Subscribe now for free; new issues every Wednesday!

Click here to subscribe now »

Want to learn more about data, analytics, and insights? Subscribe to In-Ear Insights, the Trust Insights podcast, with new 10-minute or less episodes every week.

Pin It on Pinterest

Share This