Planning For Personalization: Tips for Dev, Design, & IA

>> JON MECK:  Hi, everyone, thanks for coming.  We will be talking about planning for personalization.  Tips for dev, design, and information architecture.  My name is Jon.  I think we're ready to go.  There's not a lot of applause for anything like that because everyone's muted.  We have the participant channel, and you can say go faster, go slower, cup of coffee, clap your hands and that's how I know I'm doing okay in lieu of participation feedback.  

This is me.  This is my brand‑new LinkedIn photo.  I was excited when we took them in the office.  The VP of marketing at Bounteous.  I lead the internal efforts to make Bounteous a well‑known name.  We do a lot of great things in the marketing team.  We have a great blog post where we write about a ton of things.  

We've got so many smart people working at Bounteous and it's privilege to share that information with everyone.  

And we just love the solutions that people come up with and to share.  

About me.  My Twitter handle is there.  Feel free to follow me.  I'm a dad, digital tinkerer, make websites, go to concerts, every time I travel for work I go to a concert in a new city.  On I'm really struggling with staying at home in my lovely guest room.  

Background in marketing, analytics, and data.  Google sheets is my favorite thing in the world.  I'm ready to fight anybody who wants to debate Excel versus Sheets.  

Got a simple test if you know what you're doing, V lookup.  If you can do that, that's great.  If you've got index match, a little bit better.  And in Excel, we have the X lookup function.  There's a scale for how well and how deep we get into the functions and Sheets and Excel.  I've not tried air table but I will stick with my Sheets. 

A couple of people love to throw out fun facts.  I like to throw out the opposite.  Breakfast is the worst meal of the day.  

I am getting the participation I was hoping for.  

Tacos are overrated!  I got to tell you that.  Hand foods, I am not a ‑‑ good with hand foods.  Tacos, they're good, I'm going to have them but just a little bit overrated.  

For fun I want to tell you about fun staff.  I had never seen this many people racketing in a chat.  It's ‑‑ reacting in a chat.  It's excellent.  

I like to make number puzzles.  We got couple of books.  Sum Search is like a word search but with numbers.  

The new one, I'm really trying to break out with is ninefind.com.  Check it out while you're working from home or after you're finished with work.  

Let me tell you a little bit about Bounteous.  We have a bunch of people helping out with the MidCamp.  Big sponsor, and supporters.  It's in our backyard.  I'm in Pittsburgh, Pennsylvania.  And I am in my guest room.  So ‑‑ but a little bit more about Bounteous.  We create big‑picture digital solutions that help leading organizations deliver transformational brand experiences.  

We really like to focus on that big‑picture digital solution.  We love talking about what it takes to get to a goal.  So when we're talking about those digital solutions, what are all the different pieces that make up that process, that project.  And for us, the way that we do that, we ensure the smooth flow of data, insights, and interactions across the digital brand experience.  We talk about that, meaning a couple of different things.  We're experts many making tools talk to each other, moving data around accounts getting it to the right place and that's some of what we're going to be talking about today and it's also how we work as a company and across our teams.  We're put together with great teams of dev, design, analytics.  We've got the smartest teams, the smartest people I've had the chance to work with.  And it's awesome to work here.  

And our ‑‑ when we talk about the work that we do and how we do it, we know this is hard to do.  But that's what we do to help drive transformation at, transformative results.  

We're located all over North America.  Chicago, Denver, and Pittsburgh.  We added a Wilmington in general, and a front office in Toronto.  

Check us out.  Learn more about us.  

But today what we're here to talk about is personalization.  It's the buzziest of buzzwords.  It goes back a long time.  We've been talking about personalization forever, and really it's interesting the way that people have been talk about.  

So we can see trends from 2011, '17, 2020.  It just keeps coming up and we keep talking about it.  

So we will talk about something else.  

I'm not here to tell you that should that you should be doing personalization and doing it today.  

What I am here to tell you is that if you design and build websites thoughtful three, then personalization is easier when you're ready.  

So I'm going to give you some tips.  I'm going to talk about how we can put together a website in a way that is ready to add personalization to it.  

And this is always interesting because there's always going to be someone, usually after ‑‑ when you go to a website, you hand it off to the client, maybe they're the ones that come after you.  Maybe we have another agency that they're working with or maybe there's another dev or designer.  People are going to being with on the same websites.  

So what we have are a couple of tips in here for how we can thoughtfully put together a website in a way that go going to help everyone else after you.  Whether you're having that conversation today about personalization, you're thinking about how you're going to enable personalization on your website, or you're just trying to be a good steward and make a product that will be revered and loved by everyone else who gets a chance to work on the website.  

I've played many roles.  I've been the designer, the website builder, I'm coming here as the marketer, the guy who gets the website and adds the personalization.  I know the requests I put in and ask people to change and fix things and tweak it.  So it's easier for us to do our job 6789. 

Let's talk a little bit about the personalization.  I said I'm not going to talk a ton about it or twist your arm but let's talk about the promise of personalization.  

We know that personalization is great for a lot of differently websites.  Today we're focusing on the web but we can get into mobile, email, all of the different experiences.  But we're talking about today website personalization.  It's great for sites that have a lot of traffic.  We can collect data, use that data to make decisions.  We know that we can really do a great job with personalization. 

Sites with content.  Drupal sites with a lot of content are great candidates for adding different types of personalization there.  E‑commerce websites, sites to add personalization to and the easiest one to prove the return on investment.  Little tweaks making being difference.  You can see to on an e‑commerce site.  We tie it to transactions, customer lifetime value.  

Sites using marketing channels.  Here is the first chance to test out the participant channel.  If your doing any marketing that involves money or time, so you, someone on your team or your company is putting money into running ads, sending emails, using social media hit that raise hand button and let me know that you're doing marketing currently.  Psych a couple of people raising their hand.  Almost search doing some form of marketing, and we're investing.  When we say investing, both time and money.  

So even if you're just doing social media, you can say it's free.  It's free!  But end of the day we're investing our time.  We're putting time and effort into it and we want to make sure we're getting return out of that.  Thank you for raising your hand.  If you're doing marketing, we should be talking about personalization but we're not going to do that today.  

And we like sites with repeat visitors.  People are coming back again, and again, and that's where ‑‑ if people are coming back, we can start to build the database about people, how they interact with the website, and tweaking it to improve their experience.  

There's different types of personalization.  

So all the way on the one side, we've got rule‑based personalization.  We've got segments to create.  We can say people from here get this experience and over here get a different experience.  There's a bit of a misconception when we start talking about personalization, where everyone wants to go on the other side.  Everyone wants to talk about predictive, AI personalization, hyperpersonalization.  And there's a great article that came out last year where they were talking about how sometimes this leads to poor expectation, poor execution and people bite off more than they chew.  You need a lot of things to be firing at the same time.  You need a lot of data.  We need to check a lot of those boxes on the previous slide to know that we can make that work.  

And it might take us a long time to get to the proof of the return on investment.  

However, we don't have to start there.  And I think that's a big challenge for companies, since they read the articles, and hear the buzzwords, and they think, "I am going to do what the biggest companies are doing and I'm going to do it on the tiny website and get the same, exact result."  

Everyone should be starting from the same place.  We should be mastering the basics of personalization, whatever we're talking about AB testing or personalization, we should be getting good at the basics.  And then we should be layering up.  Levelling up, and getting to a better place.  

We can start at the early stages of web development.  We know that as an example, if we were to try to personalize a welcome message.  Hopefully search a human.  That's pretty challenging sometimes just from ‑‑ [indistinct] ‑‑ if we started this baseline, welcome human.  Welcome desktop user.  

And desktop versus mobile is ‑‑ is ‑‑ we would never think of building a website today that's not responsive.  That doesn't plan that there's going to be two different types of people coming to the website.  People on desktop, people on tablet, on mobile.  We know that these different segments exist, and we can change that experience based off of that.  

Desktop, mobile, and tablet seems to be one that we almost always saw on the dev and design side.  But a lot of these other ones are things that we can solve with other tools, personalization tools that we layer on, on top of website.  

We can say welcome Chicago visitor.  Use geolocation information to know where someone's coming from.  And we can use that to show relevant or interesting information.  

Welcome customer.  Have they bought from us before?  

Welcome, email subscribers.  Making different tools, talking to each other, and changing the experience based on that.  

Something like "welcome Drupal fan" so we can monitor what people are doing on the website and say, we think you're a fan of this type of content so we will give you more.  

So based on your previous behave we're going to make recommendations or change your experience.  

And lastly, you get to at that hyperpersonalization, welcome Jon Meck.  And there are places if you got there, you would say, that's exactly what I'm hoping for.  I want my doctor's website to know who I am.  I want Amazon to remember my last purchase.  I can check in on it.  

But there's places where we don't want to get that far.  Not every website needs to be at that level.  

In fact, it's pretty easy for us to pull up social media and find examples of where personalization got weird or creepy.  Here's an example from someone who's relearning how to drive and Amazon is recommending Student Driver magnets even though they never served for them or thought that they said it.  You can see how it might become spooky.  

And there's a lot more examples of this.  Target example, we know about.  

All of these things are probably too far down the path of personalization and possibly a sign you can go too far with this.  

Let's reel it back in and say we can start personalization with just one segment.  So let's flip back to the tips for our designers and developers and information architects.  

Sometimes we know, if you're building a website, maybe you will wear all those hats at the same time.  You're going all of the above.  

Or maybe you're playing one of those roles or working ...  [muttering]. 

So let's start with putting together a website, doing the design for a website.  This is all about flexible design.  We want a design to give options to the other people that are building a website, the other people that are visiting the website.  

And the question I would pose is, if you knew everything about an individual, how would you change that person's experience?  

Again, we can talk about personalization, say the buzzword, we know the few examples from the Netflixs and the Amazons and say, well, they're going to recommend the next in the series or a series like that, or a product like that.  

Let's think about our website ask say, if we're changing our website, we're changing someone's experience, what are the mechanisms we can use to do that?  

So if you have perfect information, you could hide the wrong information from our website.  

We can incorporate things like people's preferences.  

We can make recommendations about what we think is the next thing someone might like.  

Again, if I'm reading different articles, I can say, I think this is a similar article about a similar topic, and you might find it interesting.  

Incorporate preferences.  

And so when websites are doing this well, you sometimes don't even notice it.  An example, I logged on to CNN today and they asked me begin if I wanted to sign up for Andrew Cooper's newsletter.  If they had the information, they would say I said no 10 times and I get all my information from Twitter.  Not a good thing to admit but you can get to the news sites from there. 

We can show better, more relevant information.  

So if I knew exactly who was on my website, and I knew everything about them, I could then start to show them exactly what they're looking for.  It's getting a little bit more predictive.  You can simplify the experience.  

We can start thinking about examples of all of these things and saying, simplifying the experience, if I'm already a customer, I don't want to see a "sign up to be a customer" button.  I can just hide the things that don't apply to that particular person.  We can send the right messages.  There's lots that we can do, and while we don't know everything about everyone, let's assume that at some point we're going to go ‑‑ we're going to learn some things about some people.  

So there's the ‑‑ we know everything about everyone.  We know we're never going to hit that. 

But we're also not at the place where we create a website, we put it in the world and we're done.  

We have that website out there for a period of time.  We're often using tools that the capture information about how people are using it.  And if we use the personalization tool, we're capturing information used to then change their experience.  

So we know that we're going to learn some things about some people.  And the question would be, what do we do with that information?

For this, on the design side, we need to be able to embrace flexible design principles.  

Yes, it has to look great.  But it also has to be able change easily.  

So we say this is the ideal home page and what we've been looking for.  We should be asking these questions, saying, would it still look good if I got rid of this section?  If I knew this was the wrong section, if I had information and knew that it didn't need to show this to everybody on the website workers it still work if I just took that out?  

We had recommended articles on the home page but started putting that in more places.  On partner pages we can recommend partner articles.  

We know ‑‑ we start to learn about the people on our website and we can give them more information in different ways than we originally imagined when we built a website.  

What's another way to show the same content?  

So again, if you're creating content like blog posts or events or things like that, we've got the long formal blog posts.  We often will have some sort of summary, a card that we have.  How can we show one piece of content 10 different ways?  Again, just going back to the fact that we know that we might need to put this somewhere else.  

We might need to show it on the home page as a recommended article, or show it as the most recent article that we published.  And we want to make sure we're giving flexibility to a team that will add personalization with imagery, text, hard‑coded images and text together will slow us down.  If we need to call the design team every time we make a tweak, that's really going to put a halt to our personalization process.  

At Bounteous we love the atomic design by Brad Frost.  We have a great article talking about a living style guide and how it embraces the ‑‑ the concept of atomic design.  If you don't know about t it's simple.  Distill things down to their smallest pieces.  We can take atoms like the search button that you see here.  The one small piece but put it together and say, create a molecule that includes the title, the input, the button, say, that is our search section.  And we can put that in different places.  Move it around the website.  Put it here, put it there.  

We can create organisms like the header bar, and make layouts that we build into pages.  A quick summary of atomic design.  

But the point of this, by thinking in this manner, it's really enabling all of these other things downstream that we're not even thinking about.  If I can grab the search box and show it in a modal or somewhere else on the website, based on some information I collected about some individual, it's really going to empower your personalization teams to get things done.  

Often when we are thinking about personalization we're thinking about, what are we going to add to the website?  

So most of the personalization tools that are buzzy and trendy are client‑side.  There are more expensive server‑side.  Personalization tools.  The client side lows asynchronously, and it can swap content, add cannot.  Acquia lift is one we use.  We can feed that through content hub and lift, and swap a piece of content on the site.  Thinking about this and saying, how will we add more information to the site?  

What are the mechanisms I have to change that experience, or to add a different message to the site?

I'll give you an example.  On the top of our blog posts at Bounteous.com, we have a hidden banner we can turn on or off and we can populate with anything we want.  We build it in the website before we knew what we were going to do with it.  Before we built the object.  

We used that bar through lift or Optimize, we can populate it with our Google trainings and show it to people who come to the website from Chicago.  

Flyouts and overlays.  If we have an idea of something to share with someone, we need to have a way to put that into the website that's not disruptive.  

And there's a way ‑‑ and I will show you an example about some flyouts and overlays.  Examples of how other sites are doing this.  And then we can try to embed that somewhere in the page.  

Challenges with embedding in the page are, again, come with the client side but there's ways to fix that.  

Server side tools.  Optimize has an example of that.  

We spend most of our time on the client side, thinking about how we can use the information that we're getting from people coming in, cookies, marketing tools, and things like that.  

So we're mostly using client side personalization tools at Bounteous.  But some of the bigger websites would probably prefer server‑side for behind the science, high‑load personalization ...  

Here's a quick example of adding content to a website.  Again, on the Bounteous site we can add the breaking news strip inside of an article or just kind of pushes the content down from the top of the page.  We can have things fly in or pop up.  

And everyone hates the modal popups.  I just complain about the CNN ‑‑ I was going to use this example.  But we tried this with a newsletter sign‑up, and we saw 250% increase in subscriptions.  And there's ways to do this thoughtfully.  We're not going to spam you and hit you every time you load a page and we're also ‑‑ the way we did it, we made it clear there's a big X, and it's easy to close.  It doesn't give you a guilt factor, click "I hate saving money," and it closes.  We do it in way that's thoughtful, and trying to give you an offer.  If you want to sign up that's a great way to do it.  And we offered another couple of options.  

Email is not my preferred method.  

Great.  Here are two other options.  

And we have embedded content, inside the page.  

And you can see, we have CNN across the top.  We've got popups coming on a website.  Amazon's making recommendations.  That's just one example.  

So as we design the website, we can be planning for this.  We can be saying, how will we add information to the website?  

Or, we can add information that we know might change later on.  

So we have to say, "This is what we're going to launch with, but there may be a future state that's going to involve extra information."  

I love this motto.  

"Make it dumb.  Make it dumb first and then we'll figure out how to make it better later."  

If we're dealing with a process, yes, we can automate our sales force to this and dah‑dah‑dah, do it manually first so we can prove it's going to work and then figure out ways to make it better.  

So adding a greeting to your website that says, "Hey, welcome to the website," is a simple way to do that.  And we know we can come back later and maybe make that a little bit better.  Hey, do you know we have a sale right now?  

There's ways we can put it on the page.  Here are our three most recent blog posts.  We can add it to the website and think in the back of our mind, later, when we get to personalization, that might be, hey, here are three recommended blog posts.  

So one is a straight feed and the other will use some information that we learned about some person to make that decision.  

So those are our tips for designers, short and sweet.  Let's talk about the developers.  

We've got some really good questions and I'm going to save some of those for the Q and A.  

Let's talk about devs for a second.  Talk about how we can build it once and make it in a way that people can use it.  We're not just building to specs.  We're trying to build a system that other people will be able to use, and really the goal is to not call the dev team.  [Chuckles]. 

So let's think about it and say, who's going to be the next person to edit the website?  Often content or personalization falls on the marketing team, or a different agency.  

And what we want them to be empowered to do is take it to the next level.  

So here's a great example.  This is what happens when our team builds a website.  And then when they hand it off to me, this is what it turns into.  

We can take the perfectly organized thing that is created, and we just make a big mess out of it.  And we start dragging things and changing things.  

It needs to be some control and there needs to be some barriers built in.  

We have this conversation with companies all the time.  For personalization to be successful, the stakeholders need to develop a fast and flexible cadence.  

So I know everyone has a different time line for how quickly they can get dev from QA all the way up to the live site.  Typically a two‑week sprint might be too slow for a personalization cadence.  

One of our customers,s we meet twice a week, launch a new experiment once a week and review the results at the end of the week.  We're going quickly with a lot of these tests and experiments, and monitoring them and changing them for the next week.  You've heard stories about Spotify and Netflix.  

What that means, we can't always call the dev team.  We should be able to do a lot on the website, and there should be things, you know what?  I do need a new module.  I'm going to put in a ticket and we're going to go through the process and spec it out.  

And we want to be using the tools that have been created to tweak it, change it, keep it moving.  

If we're talking about customizing existing modules, tweaks to the site layout, swapping things around just to see if they work better.  I know in our training site, we have a big map at the top and our list of our trainings at the bottom.  That's really important during the year when we have people that are looking for training right next to them.  But now we're starting to rethink that.  We're going to need to change that where our offerings at the top and digital at the bottom.  We need to be able to do that without waiting two weeks for that to come back around.  Colors, text, hiding and showing things, make it flexible.  And we will call the dev team when we need to for the really important things, the new stuff.  

I want to call out the consider all of the downstream effects.  

Here is our superhero team.  If you think about who all works on a website.  In case anyone doesn't know, this is Marvel, the Avengers.  One of my goals is to get through all of the movies.  I have not watched any of them.  I think ‑‑ I'm up to like the third movie.  I'm live streaming with a buddy and he's seen them all ‑‑ I don't know who all these people are.  

Think about the team on your website as a superhero team.  We've got dev designers and all the people that are building the website, and there's the whole team that's going to run their website.  

And your analytics team, personalization team, marketing team.  And maybe that's just one person who wears all these hats.  

But that team needs to be able to do their jobs as well.  

So all the work that we do, if we're really thoughtful about it and we know what's coming next, we can plan for that and say I'm going to make it easier for them down the road.  

This is an interesting one, I don't know if anyone sat in on Brian's talk earlier today.  He was ‑‑ Brian Perry was talking about CSS in JavaScript.  And one of the CSS in JavaScript options he showed made really ‑‑ it was in order to control scope, it made this unique and ugly‑looking class names.  That's the opposite of what I would recommend.  In this particular scenario, when we're thinking about the downstream effects, our analytics tools, all of our client‑side tools like analytics, testing, are ‑‑ [chuckles] ‑‑ are going to use things like CSS selectors, to make it easy for us to say, "I want to know how many people are clicking on that button."  I'm going to put a trigger on there and watch to see how many people are clicking that.  

Often we're going to use a CSS to do that, and well be easier when we add unique IDs, class names that are human‑readable.  And we say, I know exactly what I'm driving with this.  

These two examples are from Acquia lift, Google optimize on the right.  They have CSS selectors.  

Acquia gives you the option to embed code on the website and say, this is the thing I want to focus on.  

If you remember from the last slides, they would involve me going to the dev team and asking them to put code on the website so that I am now able to run a personalization campaign.  And it's contrary to what we hope to do.  

If possible, depending what you're doing, we love class names nature human‑readable.  I should be able to look at that and know exactly what I'm looking at.  

Being friendly to the people you work with and friendly to other tools.  There's other ways to add relevant information to the page that other tools can ingest and use to make decisions.  

We've got things called data layers, when you're working with Adobe or Google.  We will organize this information and put it in a structured format and put it on another page.  When another tool is added, they can ingest all the information I care about.  

Think about your user.  What do we know about them?  We don't want to put in things like their name or email or social security number.  That's ‑‑ that's not what we're talking about.  If you know they're logged in and a user, a former customer, that information can be added to the page and we hand it to analytics and they can adjust the information and track it.  Or the personalization team, this is an example from opt mice ‑‑ they can say, I want to show a particular experiment or personalization to anyone that's on a blog post tagged with Google tag manager.  

So we can be friendly with the other tools as well and make it easier for everyone else to use that information.  

Modular design.  We talked about it on the atomic design side.  But if we can build it into the website so whatever content editor ‑‑ so whoever is the content manager, the marketing, it's easier.  We want things to be reusable, self‑contained, and customizable.  

And further down the line, if you are going to start thinking about personalization, your dev team should be involved.  We know they're going to be involved earlier, that's when we can start talking about unique things to personalization.  One of the challenges with client side is the flicker that might happen whenever a client side tool tries to load as a page is loading.  

Ask it ‑‑ make sure your asset is loading in the proper order.  Think about the QA and the sprints, and whether or not personalization should be a part of that. These are things to talk about as a team.  

When we think about team, again, when we think about the personalization team, we usually try to make a team that is built up of dev, design, marketing, and analytics.  Those are all of the key stakeholders involved in the ability to launch a personalization, the intent behind some sort of personalization campaign, the measurement of whether or not it's successful, and whether or not it looks great.  

So a lot of things from the dev side.  Again, if you follow all of the advice, all of the instructions that we're putting here, even if you don't do personalization at all, the end result is you have a really great website.  All of this is being thoughtful and purposeful in how we put it together, specifically in other ways that help other people who then try to get to the website.  Let's talk about IA.  We've got so much information on the website.  I think one of the questions here, is it important to use data to inform content creation, content targeted.  

There's a lot of different ways that we can use information, use data, but one of the first things we should do is capture it.  

When we talk about the ‑‑ this is my tag everything ...  

Get it??  

Again, give me claps on the channel.  I'm missing out on this audience participation.  

This is ...  

Okay.  

What do I know about the pages and the content on the website?

Okay.  When people are creating pages, they usually come in with some sort of thought process behind it.  

We want to make use of taxonomy.  

We want to use as much as we can to force people to classify the content on the website.  

I would say most people are really not doing this, unless someone else is asking them to do it.  

We can be that person.  When someone hands us, you know, some contents, "Put this on the website," we can force them to say "Before we do that, we have six fields to fill out in order for us to quantify and tag it and make the information available."  

We want to overuse a tagging structure that creates the content and communication the substance and intent to the other tools you're using.  

So simple, most people know about tags.  Keywords.  Almost every piece of content can be tagged with different items.  There's themes, and products or something more specific.  If you're selling something, there's the specific thing you're selling that usually belongs to a category or a theme.  

In this case, thinking about writing blogs on the Bounteous website.  We might be writing about personalization but we will tag all the different technologies we use.  There's two different sets of taxonomy and tags that can be put into the data layer, shared with our other tools, things like that.  

This one's easy.  I think a lot of people are using keyboards already.  You see them at the top of blog posts, click on it and sort someone's content library based on these tags.  

So we're building in a bit of functionality.  We're building in a bit of logic, and the ability for people to group things together.  

So this is great for the experience on the website.  This is also really important for that other stuff.  Learning what people like.  

I can't know what they like unless I can ingest that information and make sense of it.  Tools like Acquia Lift.  They can take in the keywords and say, you rate six Drupal blogs and only three Google blogs so we say your favorite topic is Drupal.  They do the work if you pipe in the right information.  That's why we want to talking about these things early on in the site‑building process.  

Who is this content for?  As we write this, often you have the idea in your head or the author's thinking about it.  

We're doing this for a very particular reason.  We're hoping to hit a very particular audience.  

Track that.  

Put that into some sort of taxonomy.  Maybe someone choose from a dropdown and say, it's got to be one of these five.  And that will help us put that into other places.  And we can say, we think this person fits into this model based on what they're interested in, what they're reading on the website.  

And we've also got sections of the site, content types.  This one is interesting.  

We can understand what type of content it is based on where it's located.  

So our service section, blog section, about us pages, we have all these different sections the website.  It would be great to see that information.  Know what sites people are ‑‑ what sections people are getting the most, or what particular content type is being consumed.  

When people say things like, "I have a ton of content on my website," we want to say back, "What's the most valuable content?  How can you prove that?  How can you say this is more valuable than that content?"  

There's a couple of different ways to do that.  Google analytics has a built‑in page score that will take the value of a session and spread it out over all of the pages reviewed before that. But often we have a bit of an internal idea about how valuable a particular page is.  

So for instance, if I have a website, and I am trying to ‑‑ I want people to read blogs, that's a great thing to do.  I want people to do that. 

I also want people to check out my services and know that I build websites and do all these other things.  That's a more valuable page.  

When we say how valuable, we can create a bit of a relative scale.  And we can say something like, "Our services pages are worth twice as much value than a blog page."  We know someone is at this section of the website or on this particular page, they're a little bit more valuable to us than what our eventual goal.  

If we're trying to get people to fill out a form, there's indicators.  

So content scoring is a super awesome thing to do and we can do it right.  And it really forces our teams to think about how we're organizing the content on the website.  

We did a fun project on one of our websites where we could measure how ‑‑ how many people went from a low score page to a high score page.  So this content of uplevelling people and saying, we had a ton of traffic to the blog but we need to get more of those people to go to the next section, to invest a little bit more time or read a little more critically.  And we can measure that and say is that working.  

Here's an example from Lift where it tracks it and scores people and at the end of the day we can take the list and say, "Show me the most valuable people on the website, and let's change their experience a little bit."  

So we know that based on their score, these people are only ever going to read blogs so we should be giving them the best blogs and more blogs, and all they want is blogs.  But these people have the potential to go to something else.  

Maybe we will hit them with a little bit of a different message.  Try to send them in a different journey.  

We talked about design.  We talked about dev.  We talked about IA.  Let's talk about what's next.  And we're getting to the end of time.  We're doing pretty good.  

Personalization.  We said it's great for sites with a lot of traffic.  Sites with a lot of content.  E‑commerce sites.  I'm going to come back to the sites that use marketing channels.  

I did a webinar with Acquia last year, Eric and I talked about this.  And saying what people tend to do is get stuck in the ideal, or the vision of personalization.  

Again, it can be a little bit messier than that. 

We want to get people in there and starting to use personalization tools.  

Get in there.  Start seeing what it's like.  We can do a fifty‑fifty split and say these are east coast/west coast.  This Pittsburgh audience will get a different experience.  Marketing channels is great because it tells us how someone get to the website.  If we're using best practices in marketing, we can say, these are people that came to the website through Google ads, and they clicked on this campaign so I know what they're interested in.  The words they clicked on.  When they get to the website, I should make that as relevant as possible.  

If I know they're coming from my member newsletter, and already subscribed or purchased, why should I hit them with a sign‑up message when they get to the website?  In I'm tagging my content and they click through the email and land on my page, look at the URL, and I know how they got here, and make the experience match the message they saw on the inbox and the social media, a Google search.  And marketing channels are the first place that we can start with personalization.  Because we're already spending time and money on this.  

We can just carry through that experience from outside our website to our local site.  

And I was going to pitch you too hard on personalization, but two great tools to think about, Google optimize, that's a freemium model.  It's easy to set up and get it loaded on your website and start trying to use it.  

Acquia Lift, we love it with Drupal sites, the way it feeds content is amazing and makes it super easy for us to change experiences on there.  

And with all of that, the whole point of today is let's design, build, and architect websites thoughtfully, better thinking about what comes next.  

And if you're handing off a website to your stakeholder, your client, and saying, "Hey, I built you this website, it did this for you, and you're ready to go, you're signed up, you can take this and immediately start doing any sort of personalization you want on this, because we were thoughtful in how we put it together."  That might be a great segue for you to do that for your client.  That might be a great chance for your client to take that and run with it, or maybe it's going to be handed off to someone else.  But if we think about it, if we plan for these things upfront we will safe ourself a lot of time and effort when we start to get these requests later on.  

And here's my personalized response based on me.  Thanks yinz.  That's how we say y'all or you guys in Pittsburgh.  

Thank you, any questions?  We've got time for Q and A.  I'm not going to miss the opportunity to plug the contribution day, this Saturday.  And we have a feedback form.  And I will leave it here.  If there are questions, I'm happy to take those as well.  

Thank you all for coming.  

I'll give you guys a clap in the panelist channel, because you guys are awesome.  

Questions?  You can unmute, chat.  There's a lot of different things.  

>> ANDREW OLSON:  Thanks, Jon.  What I would like to do is check the chat.  

And then I can unmute if there's a question.  

>> JON MECK:  I'm got a question.  The question is, is it more important to use data to inform content creation or should content be targeted?  

This is kind of a ‑‑ everything is an "it depends."  There's not a lot of perfect answers in marketing.  

I would say that content creation should match what you're trying to get people to do.  

So you're trying to eyeball, people coming to the website, the perfect way is to start mining your data and saying, what works, what gets people here already and spending time on.  And that's the type of content you continue to produce.  But sometimes you write content that doesn't do well but that means you're going to use it in a different way.  If this is a piece of content that ties to my mission and vision, or I know it's supporting a particular service or offering that I'm trying to get up off the ground, I'm going to have to use this content, send an email or show on a home page, do something extra because it might not get that organic traffic.  

It can be double‑edged.  If you only follow the data you will miss out on some of your quarter business objectives and then end up incorporating some of the bad practices that we don't like people to do, the clickbait articles and things like that. 

You want to be smart with your data.  Track it and understanding it.  It's easier if you take two different types of content and compare, how did this type of content do?  

Can I prove that long articles are better than shorter articles?  These are similar.  Trying to figure out why one did better than the other.  

I hope that answered your question.  

>> ANDREW OLSON:  I will ask you to raise your hand and I will unmute you.  Otherwise, Jon, you're doing great, answering the questions in the chat.  That's working well.  

>> JON MECK:  David is asking ‑‑ I would love to hear more about the success criteria beyond "this thing appeared for a user."  

I think you're asking how do we measure how this is working.  And the Gardner article I referenced said depending on the site you have, you may need to change your expectations.  If we're doing A/B testing, it's easier for us to put something at the end and say, "This is what we're trying to fix, two ways, and we want to know which one is better."  

Personalization, if you have the big site, a lot of content, an e‑commerce website, we try it back to a discrete metric and say we're trying to increase this and do it in this way.  

But often, personalization is last also used just to improve your brand experience or your loyalty.  When I go to Amazon and it says hello Jon, it's not going to make me buy anything more but it has a bit of a connotation to it.  

We try to tie it to marketing.  We do a little bit of A/B testing, control groups and say we will send a campaign in here and show some people the personalized experience, and some people ‑‑ not the personalized experience.  Sometimes you do personalization because it's the right thing to do and it will improve the experience for the end user.  

Especially when we start talking about things like hiding the wrong information, or respecting people's preferences.  Those are things you should be doing anyway.  We just have greater capability now with the tool that the allow us to change that. 

Maybe we can chat easier, curious about server side versus client side.  

One of the things I love about client size personalization is the flexibility that we have in the client side.  The ability to go and launch a change to your website in minutes.  I know with the coronavirus, we could see how long it took certain people to update their website, change their messages, to do things like that.  

It could be really easy for someone to go in and launch a popup or a banner or a message, change their home page, or hide their event schedule or anything like that.  In minutes, depending on their setup.  

So the client side tools are often cheaper.  They're often easier to use, and there's a little bit of a lower tech curve to adopt them.  Server‑side is usually bigger companies that are staffed.  

Anyone else with questions?  

Again, the point of today isn't that you should all be doing personalization.  Not every website needs personalization.  But the qualities that we look for whenever we're adding personalization to the website, we go back and we say, what's available for us to use?  What sort of data do we have?  How are we going to be able to grab items and move them and change them and tweak them?  And we're often ‑‑ in the code, and we're looking at classes and ID names and we're saying, "I can't use this.  I need to do dev work in order to get started to do the personalization work."  

Lauren has a great question.  How do you ensure personalization isn't keeping people in their bubbles?  Like, "I noticed you came from a makeup website so you probably aren't interested in articles on engineering."  

That's a great question.  I would say when we think about how other sites are doing this, we think about how we can add personalization.  

There's sometimes the ability to swap content.  I mentioned the example where the three recommended articles we could swap that based on what they want.  There's a couple of fall back methods and a couple of different ways you can do that gradually.  

So probably a better way would be to leave the default options there and add another section beneath it that said recommended options.  You will see that on Amazon page or most e‑commerce sites there's five different boxes.  Recommended for you, people similar to you, but we think we should do this.  Our partners have paid for us to put this here.  We can see the different options and measure which ones do better.  Google analytics has a free tool called enhance the e‑commerce.  That says we can track the third item and the second recommendation zone, that's your most popular one and that drove the most pull‑through revenue.  We can track and measure which ones are working.  

I think that's probably more of an internal thing when we talk about what is the intent of the personalization, when we do a campaign.  That's why I address your question.  We need to be thoughtful and aware of the bias we might put on top of it.  

Can the personalized pages still appear in organic search results ‑‑ [muttering] ‑‑ sorry.  I missed the beginning.  

Any idea how Google charges AJAX‑loaded pages?  Is the crawler able to run and parse JavaScript effectively now?

We are still ‑‑ Google's going to ‑‑ most often Google will grab the ‑‑ the core content on the page.  

So it can still ‑‑ if we have like a JavaScript framework where the page is loaded after the fact, it can still handle that.  It's going to execute JavaScript and see what the page is.  But what it's not going to do is take one of your variations or personalized pages.  

The organic should be the basic, and most pleasing to the most number of people.  Could be linked to a campaign or based on observed behavior or a number of different signals to put together to change that experience.  

I'm not sure if I'm getting exactly to your question.  In general, Google's going to show you the base page and not the personalized version of the page.  Also because there would be 18 different versions or 100 different versus of how that page can look, based off of a number of different factors.  

When we come back to the makeup example.  If we had a personalization for makeup loving engineers, that wouldn't be discovered through organic?  You want different variations.  If I had one page with a couple of different variations for makeup lovers who like engineers, makeup lovers who don't like engineers, makeup lovers who like football ‑‑ by creating all those personalizations, you're not going to five different index pages that all resolve to the same page but have different experiences.  

This has been super fun.  I think the Zoom thing is pretty great.  I don't know how you guys think about it you by the fun to watch people collaborating and the speakers have done a great job.  Thank you all for coming and chatting.  I'm happy to stick around.  I don't know if there's someone else after me.  I'm guessing there's someone ‑‑

>> ANDREW OLSON:  Actually, this is the last one, so we can hang out a little bit more.  Keep the questions coming ‑‑ 

>> JON MECK:  I think it's just turning into angry people now.  People not happy when I talk.  [Laughs.] 

They hope I had better ‑‑ [indistinct]. 

I think the travel to the southwest is what we've noted.  When that's an option, I'm going to head there and get some tacos.  I did have a great breakfast taco in Austin.  I will tell you that.  

What do you think, Andy?  

>> ANDREW OLSON:  I'll let you call it, if you can take more hits on tacos.  You've done really well.  It's your call.  

>> JON MECK:  That gets in the recording and I'm going to stop the recording right now.  I'm happy to hang out.