Podcast Description
Podcast Title: Digital Rage
Host: Jeff Byer
Guest: Geoff Atkinson
Summary
In this episode, Jeff Byer and Geoff Atkinson deep dive into the more technical topics like dynamic rendering AMP, edge delivery and structured data markup. Jeff shares technical techniques he uses for his clients to help with their page speed and overall success rate, and he admires the way Huckabuy software products work and how they get results.
What You'll Learn in This Episode
- How Huckabuy Cloud and dynamic rendering work
- How edge delivery works and possible benefits
- The advantages of using a paid software solution over a free plugin
- The benefits of using an automated structured data software solution
Time Stamped Highlights
[10:04] Huckabuy Cloud: How It Works
Jeff: Explain the cloud rendering portion of your product.
Geoff: Yeah, for sure. There’s this concept known as dynamic rendering, which basically means that sites load dynamically based on what is calling them. The simple example is, if I go to a website on my mobile phone, I'll get one experience. If I go on my desktop, I'll get a slightly different experience. And that's all well and good. It's best practice and Google supports it. What they've really changed though, is they said, “Well, now you can actually give an experience just for us.” And that's opened the door for us to basically create Huckabuy Cloud. Now, you still have to have all the same content, you can't you can't do any sort of tricks. You can't keyword stuff or you know, alter the page in a way. It still has to be the same content and information, literally the same site as the user experience. But you can now queue up what we call “Google's perfect world.” So we do much like what you're doing with Gatsby. But it's a little bit different level.
This is how Huckabuy Cloud works:
- We take a site that might have quite a bit of dynamic content (whether it has dynamic content or not) and translate it into a flat HTML version. That flat HTML version is usually 30% on average of the size of the previous page, and if you actually look at it, it looks identical. You know, the way that it actually looks and feels is identical, kind of shows how much code bloat is actually out there on these websites.
- We then take that lite version and host it in a caching layer. We have a partnership with CloudFlare to provide edge delivery for this HTML version.
- We then obviously add structured data, world-class structured data to the top of the page and sort of queue up this perfect Google crawl experience.
This is a product that SAP uses, and a lot of enterprise customers that have very complicated sites with lots of business requirements are now able to sort of have this fantastic crawl experience.
One of the things that we've noticed that's interesting is a lot of a lot of websites have more index issues than they actually know about. They have a very big site that's relatively slow. We find a lot of their content, even though they'll have a big content strategy, it actually isn't even getting indexed correctly. So Huckabuy Cloud frees up all those issues. It takes care of a lot of issues and really allows a complicated website to get properly crawled, indexed, and, you know, get sort of the search attention that it deserves.
[12:39] Dynamic Rendering: Is it Related to AMP?
Jeff: Is that related to AMP in any way? Are you serving AMP pages?
Geoff: You know, that's a great question. This product actually started as an AMP product. So our first intention was, “Alright, let's build you know, have it have a sort of automated AMP version of the page.” Then we kind of figured out that Google, despite them saying that index and crawl any type of AMP page, only has certain bots are really the only ones looking for AMP pages. So when you're SAP or someone like that the type of bots that are hitting that site actually aren't looking for amp links. So it almost kind of made the product look a bit useless. So then we got into sort of the dynamic rendering world. So it does not create amp versions of the page. I think, you know, as Google includes more and more, you know, page types in their amp index, that could be a very viable product. But for now, it is much more around just building the flat HTML version and this perfect crawl experience for Google, rather than AMP pages.
[13:48] Why AMP May Not Be Adopted Across Industries
Geoff: It'll be interesting to see sort of how AMP plays out and how much you know, because developers hate it. Google, obviously, loves it. It's very strategic for them to be kind of hosting the internet. So it'll be interesting to see how AMP ends up playing out across industries. I think obviously, it's been somewhat of a success, or at least it's very well adopted in the news and publishing arena. But to see if it actually gets into the mainstream and other industries, like B2B software, e-commerce, I think kind of remains to be seen. There's just so much functionality that's stripped out that I think you'll have a lot of business users sort of pushing back on whether AMP works for them or not.
Jeff: Yeah. And so the other thing about serving a site that's not dynamic or serving static HTML, is that AMP can make it AMP by just removing the tags that they don't want. Google has the power to do this on their own. But their first attempt was making the developers convert over to it and that obviously hasn't gone very well. It goes well for new sites and people who want their content scraped and served elsewhere. But yeah, I did not adopt AMP because none of my sites are really new sources. We have content strategies, but AMP wasn't wasn't a good solution. And, you know, every report I see online from an SEO perspective is that AMP does nothing for their rankings.
Geoff: Yeah, I think it has been a bit of a failure outside of the news industry, they certainly could do it themselves. Although they typically take this approach of accessing developers first. Think about their Chrome and Chromium product. The way that they render pages and the way that they've sort of taken over the way that the internet is literally rendered and served through Chrome browsers. They're the ones that get to decide whether flash, for example, is the viable language or not. And they kind of did that surreptitiously. I don't think many people realize how powerful their takeover of the browser world via Chrome and the way that they render using Chrome and Chromium really is. So they typically do go in this way, you know by trying to get adoption within developers, I just don't think they expected so much pushback. They made it so rigid and, and non-functional that they upset — probably yourself included — developers that are coding pages and creating websites, they're just like, “Nah, this doesn't work for me.”
Jeff: Right. If we could easily dump out our current content into an AMP page and serve it that way, that'd be fine. But there's so much else that goes into it, because with most of my clients, the goal is some sort of interaction or some conversion action. That's not possible.
Geoff: Yeah, they need an add-to-cart, a form fill out, exactly. They need some some, some sort of interaction with the page that allows a customer to, to convert and that an AMP is very difficult to do any kind of work like that with.
Jeff: So your clients they’ve got some pretty large sites and some dynamic content I'm not. So from what I gather is that you're scraping the content and putting it in a more packable format for Google and serving that on the cloud?
Geoff: Yeah, so we actually have our own renderer. So it's a little bit less about scraping and more about actually rendering each page, converting and stripping out non-important aspects of the page that Google doesn't care about, and converting anything that's dynamic into flat HTML. And whether it's dynamic or not, to be able to sort of lighten up the page, make it a lot faster, and then host it using a partnership with CloudFlare, edge delivery, just has great performance aspects to it.
Google just kind of can cruise through a very big site in a very short amount of time. And that's really what they're looking for that's really ideal for them. If you think about how much time, resources and money that Google is spending, waiting for pages to load for them to be able to understand websites, it's got to be infuriating. How many times does Google have to yell from the mountaintops that page speed is important until people start listening? Generally, their main motivation, I really don't think is actually the user experience. It is when it comes to mobile. You do need to be fast for mobile to really work, especially as they grow and really their growth, you have to have that fast page speed because the connection speeds are low. But they're selfish reasons are that they waste a ton of money waiting for sites to load. So it's like having the lights on but no one's home. They're just sort of sitting there waiting for a page to load when they want to be downloading, indexing information and content. So it's just probably frustrating for them and that's why they reward sites that are fast, because they're able to access the information in a timely way.
Jeff: Yeah, it makes their bots more efficient
[19:08] Dynamic Rendering: Cloud-Based Sites Don’t Affect Canonicalization
Jeff: So, how do you handle canonicalization across cloud sites?
If there's a front-end facing site, and then this cloud-based site, is Google able to differentiate that or do you have to canonicalize everything?
Geoff: So you actually can tell them with the server response. Google will come in, they'll say “user agent is a Googlebot,” and we’'ll recognize that and say, “Okay, this is the version for you to come crawl,” the way that they kind of can just verify that the site is accurate and the same as the user experience is stuff using like their evergreen bot in their their JavaScript bots that will come in and actually look at that dynamic content.
That's how we theoretically think that they can check the actual user site to make sure everything is matching up. We don't actually see them doing it very often. They really do enjoy having this dynamically rendered version. All the communication of the bot, when they come to make a request, is done server-side through. We talk to them at that point and say, “There's a Google friendly version of this site.” And that's what they go into and crawl and I'm sure they can, if they want, hit the actual user site. They have many ways to do that. But we just don't see them doing it very much.
[22:03] Edge Delivery Helps Improve Page Speed and User Experience
Geoff:Edge delivery is just saying that instead of having to call a server, and that server responds and works through all this stuff, there are actually distributed servers around the world that are based on locations that each have all the information that's necessary to serve a page right there located in memory. So that, say, a bot comes from China, it's still going to be instantly available in China and it doesn't have to make the rounds around the world. It's right there available at the edge. And so this partnership with CloudFlare is exciting for the implementation purposes, because it's just a very simple DNS change, and then actually, they don't just get Huckabuy Cloud, they actually get 30-50% improvement in page speed. Which is pretty exciting.
Jeff:Yeah. So we've been implementing CDNs for a while now. And it does help a lot. I mean, there's pluses and minuses, the implementation and the and the distribution of assets to the servers, which, you know, you want to make as seamless as possible. But yeah, we do understand the benefits of a CDN, especially on sites that need to perform internationally.
Geoff: Yeah, I mean, just listening and researching what you're doing. You're doing it exactly right. I mean, you think about page speed and delivery and how to sort of have a site be optimized for what Google wants. We always say, “Google is actually pretty open and honest about what they're looking for out of these websites. And if you give it to them, you get some real nice benefits.”
Jeff:Yeah, exactly. Yeah. And we will monitor all that for our clients. Our clients are usually small to medium sized businesses. We do have a couple of medium to large size international B2B manufacturers. But for the most part, we we try and keep it simple and make sure that what we're putting out into into this the web is, is very clean HTML structured properly and has clear signals to Google structured data, everything like that to what it is what it you know how to use it, how to crawl it, and who's responsible for it, which is now a big thing with the E-A-T.
Geoff: That's fantastic. And I hope your customers understand what they're getting because it's pretty rare that even in the enterprise size or small medium, it's pretty rare to have someone be able to provide that and you're obviously doing that. So I hope they appreciate you.
Jeff: Thank you and my clients have absolutely no clue. I keep the tech-y stuff out of the conversation.
[24:53] Technical Mechanics of Huckabuy Cloud: How It Works and Product Features
Jeff: So do you mind if we get in the weeds a little bit a little bit of how the product was built? What kind of development stack are you using to implement this?
Geoff: So the main application layer runs on Google Cloud, which is a phenomenal product. So just for warning, I am not a developer, you're more experienced I am. I happen to have a fantastic CTO, who's brilliant and really has built this thing. I know that we've almost finished my version completely onto Google Cloud. Google Cloud is incredible in terms of what it can do. I think it's a real disrupter. And it's going to sort of take down AWS and a bunch of other these sort of storage and application layer providers. So yeah, almost the entire application runs in Google Cloud. And then obviously, our edge delivery, and how we're actually serving the site, is done through this partnership with CloudFlare, which is Also is just amazing.
So, you know, my background goes back to Overstock, and in the days where storage was expensive and just the amount that we spent on tech was kind of just for storing things and serving the site was just sort of mind boggling. I remember we had like, server farms in Utah here. And to think what you can do now is just sort of mind blowing. A lot of these technologies, too, aren't really getting leveraged yet like CloudFlare and edge delivery, we’re one of the very first if not the first partner sort of leveraging this for actual business reasons and intentions. And the when I every, you know, every week when I am one-on-one with our CTO, and he is describing what Google Cloud is capable of. I'm just like, “This is insane, and their prices are so low.” And so yeah, almost everything is between Google Cloud and CloudFlare.
Jeff: Okay, yeah. And we've been experimenting with a lot of different cloud based services as well. So in Google Cloud is your main application and the client sends the DNS through the application, then the application is able to do all the translation in the cloud?
Geoff: Yeah, so it does the actual rendering of the site into an Huckabuy Cloud version of the site. And that's done through Google Cloud.
Jeff:So the proprietary portion of this is how the app decides what is required and what is not?
Geoff: Yeah, so it's the translation of the page and the conversion of the page from whatever it is today into that light flat HTML version. That's really the big chunk of the proprietary technology.
Jeff: And so does it do the same thing, like on a lighthouse report, when you see that you have unused CSS, is it something similar to that where you can actually strip out unused CSS?
Geoff: I believe so this is getting a little bit more technical than I'd know about the product, but I believe, yes, it's includes CSS that's being used, that's actually, you know, serving the helping the page load, look the way it does, but unused CSS and things of that nature, I'm quite certain are being stripped out and taken out of the page.
[28:28] Huckabuy Cloud Works on WordPress and Across Platforms
Jeff: Okay, and so, have you handled any, like, you know, having to translate through WordPress or any other of the major open source platforms?
Geoff: Yes, we're pretty platform agnostic, because we're just looking at the front end.
WordPress is actually quite a guilty party when it comes to page speed with the amount of code bloat that goes on. I think it's a fantastic platform. We're on WordPress, and a ton of software companies are on WordPress just because of its flexibility, it's relatively SEO friendly, there's a lot of nice plugins and such, but we do find that a lot of WordPress sites have a lot of these issues where there's just slow page speed, a lot of code bloat and oftentimes there’s a lot of dynamic stuff going on. So, Huckabuy Cloud can help WordPress sites quite a bit.
Jeff: Yeah, I've been experimenting with accessing the WordPress through the API, and translating the API server side, so that it spits out static HTML. And right now, the first implementation I have is through PHP. So it uses the PHP connection to render it server side. So it does cause a little bit of a delay. So my next thing I'm going to try is doing a server side cached solution so that the content is always there and it only updates the API when the site changes when something's saved. I’m trying to give the client a WordPress back end that they're familiar with, but serve ultimately static files on the front end.
Geoff: That's brilliant. I think that's a really great way to approach it. We haven't gotten to the level of like actual integrations with CMS, which it sounds like that’s where you’re headed, to where the user can actually still interact with the CMS of their choice, but they're going to get this really nice quick site. We're more taking whatever that CMS spits out and then translating that into a faster site, but I like the way that you're thinking about it.
[32:31] Huckabuy Success Stories
Jeff: What are some of the big success stories that you've had with your software solution?
Geoff: Yes, if you think about just structured data in and of itself, that product has gotten us a long way. It's gotten our customers a long way. So our average customer that's using the structured data product, and this is funny, I remember Matt Cutts always talking when unstructured data first came around about how structured data doesn't influence rankings or doesn't Influence traffic, but our average customer just on structured data product grows 62% in 12 months, which is pretty incredible.
So structured data definitely moves the needle, if you just think about how much Google understands, with a site with very good structured data versus a site that doesn't have structured data. Think about like an SAP, you know, here's a really complicated website. It's not organized like an e commerce site. There's not product pages that are all structured, there's not category pages that are all the same. It's a very complicated product, very important site, very high domain authority. When they come to a site that's like that, it's relatively hard for them to figure out what's going on. So if you layer structured data, like World-Class structured data, across the entire site, you're saying, you know, “this is a software application and then integrates with this and there's the price,” you know, those types of clues. Google wants that information so badly. So if you go from SAP, it's a very dynamic site, to where you're actually sort of spoon feeding Google all this information per page via structured data, the amount that they understand about SAP jumps through the roof. So we have these really great success stories.
[36:39] Structured Data Markup Software Implementation and Features
Jeff: The structured data service that you offer, is that somehow automated or is that by hand?
Geoff: That's automated. Yeah. We populate a JSON-LD packet of structured data across the entire site as they change content or add products or whatever it happens to be, the structured data automatically picks up the new information.
As Google also changes their requirements, we do have to make a change when that happens, but it happens really quickly. And it goes across every single customer. So 50+ customers say, there’s an event markup change, just like last month, everybody's fixed within 24 hours, while the rest of the internet kind of lags behind.
So, yeah, the automation of it and the depth of our structured data is really the differentiator. It's very in-depth. It is automated. So it's a big box to check for sites to just not have to worry about it anymore.
Jeff: Yeah. So we've been doing all of our structured data by hand and I can tell anybody who is a developer or who is starting to look into structured data on your own, the schema keeps changing constantly and they're adding more categories, and then they're adding more identifiers to each of the categories and you can spend months going down this rabbit hole of schema, this schema, that schema that you know everything. So having an automated solution, that is able to recognize what the content is definitely valuable. So I see a really high value in that product.
Geoff: Yeah, thanks. We think of it as something that lends itself to outsourcing because it really is almost like a full time job. If you want to do it really in-depth, in-house, you almost need a developer working on it almost all the time. And that's just difficult to even find a developer that cares enough about structured data to want that job and to do it is a hard thing. So the automation piece really helps our customers for sure
[38:42] Even Though Structured Data isn’t Technically a
‘Ranking Factor’ It Will Improve Your Site's Ranking: Here’s Why
Jeff: Back to what you mentioned about Google saying that structured data doesn't help with rankings. So they'd say that because it's not a direct ranking factor. What it is, is you telling them what your content is and that content ends up being your ranking factor. So they always say that this is not a ranking factor, but there's things that are ancillary to what you're doing that are. And so structured data is huge, because all you're doing is providing information to Google that they maybe previously didn't have. And that exactly increases your rankings.
Geoff: That's a great way to put it. One thing that we see happen, very quickly, is you start ranking for a lot of new keywords, and specifically mid-tail terms that are sort of like the bread and butter of any great SEO campaign. You start making positive and authoritative connections to these mid-tail terms that just weren't there. You know, even if the keyword was in some sort of metadata or HTML, it's not authoritative. Metadata is really suggestive but structured data is authoritative. So the connections that are made to new keywords, are average customer in 12 months there number of ranking keywords in the top hundred of Google, on average grows 91%. So, that's really the first step is that you just get these positive associations with a whole bunch of great keywords that just weren't there before.
Jeff: Yeah, it is a great way to expose different content to different audiences and find those low funnel keywords that are directly related to your client and your sales.
[40:22] SEO Tools Huckabuy Uses
Jeff: So just finishing up here, what are the tools you use on a day to day basis?
Geoff: One thing that I find that we find very valuable is — so we do a lot of keyword research, which I think is so important and often so overlooked. Just the importance of keyword research, not only for your SEO, but to learn about your business and what people call things, and how they search for things, we do a lot of like navigation and optimization recommendations with our customers outside of the software, and nothing beats the volume numbers and information within Google's Keyword Planner, because it just comes straight from Google. We use Ahrefs a lot. It's a fantastic tool. We use Google Keyword Planner, because I find it to be the best at suggesting new keywords in actual volumes and then kicking it over to a Moz or an Ahrefs to get the Keyword Difficulty scores. We also leverage Google Search Console API.
We have a dashboard that's in beta right now that's pulling in a lot of statistics from Google Search Console and from Ahrefs and actually, as they have sort of minimized and gotten rid of the old Google Search Console, and kind of don't show as many crawl stats and crawl information, we actually are the loan company that can provide that information, because we're actually monitoring the crawls and what how Huckabuy Cloud is interacting with Google, which is pretty exciting. We never really did a lot of dashboard type reporting. And that's a new product that's part of our product now that people can log in and see really cool SEO statistics that are kind of going away.
So we’re leveraging some cool API's and SEMrush API, Ahrefs, Google Search Console and then CloudFlare, of course, those are the hot tools around Huckabuy right now.