 Hi, guys! I have to be honest, I was a little upset when I found out I was going after Tom just because I have a really strict rule. I don't like going after people from the UK because they sound really smart and I sound like a Kardashian, but we are going to roll with it. So my name is Jackie. For those of you guys that haven't met me, and today we're going to talk about Enterprise SEO. So I think the first question we should ask ourselves is, what is an enterprise? What counts as an enterprise site? So per Google, which hopefully most of you guys know how to use, an enterprise site is anywhere from about 250 to 1,000 employees minimum. Luckily, that means I have worked at some enterprise companies. So I've worked at Square, also known as Block. Sometimes people call it Jack Dorsey's other company. Uber, Dropbox. And now I am back at Uber. And needless to say, I love working enterprise sites. They're a little messy, they're a little crazy, but they're multinational. They have a lot of services. They have a lot of product lines. It's just really fun. And if I'm being really honest, I also just hate building links. I be really happy to never have to build a link ever again in my life. And the nice thing about working in an enterprise site is you already have the signals. You just need to make them a little bit more clear for Google to understand. You already have the product market fit. You already know that your product is good enough. You don't need to fake those things. And the other thing is I love working with people. If you guys can't tell, I'm actually a people person. And the reality is most enterprise SEO issues kind of tend to repeat themselves. A lot of big enterprise sites, like I said, they're all really the same. They have internationalization issues. You just have a lot of services that you have to make up to part. And they're really being blocked instead by people issues. Elita once asked a group of about 200 SEOs. She asked them, you know, what is the biggest blocker to you getting your SEO initiatives actually over the line? About 43% said it was a lack of resources. 25% said no platform flexibility. 21% said a lack of management support. And 11% said it was a lack of comms or coordination. But the reality is about 75% of those most replied replies, those are all really fundamentally people issues, right? Like brand is blocking you. Legal says you can't say that. Those are not really SEO problems. In a small company, you can really do everything yourself. But if you really want to win an enterprise SEO, you just have to really nail these three things. The first thing is you want to identify the most impactful projects. The reality is you're never going to run out of things to do for SEO. There's never ending list of work that you could possibly do. So you want to make sure that you're identifying the most impactful projects. Two, you want to ruthlessly prioritize and forecast to get buy-in and get really, really, really good at storytelling. Number three, you want to over communicate your wins. Say them over and over again. Send them in newsletter. Send them in emails. Put them in your decks. Put them in planning. It's really important. And make sure you screenshot everything. So the first thing we're going to talk about is identifying issues. I'm not going to talk about this too much because I honestly, in my heart of hearts, think that most of the time, you know what the SEO issue is. I very rarely meet an SEO in an enterprise, and they literally don't even know what the SEO problem is. But in case you want a little bit of a reminder, here's a dump of the most common enterprise SEO issues you'll find. So the first one is things like crawling issues from features. Maybe you're bringing in context from a user interaction. You know, maybe it's coming in from like a scroll or a click. Google does not scroll. It does not click. Things like frameworks. Maybe your company made your own framework that blocks bots. I've been there. It's not fun. Don't do it. Maybe you have a duplicate content issue. This is actually pretty common in the enterprise. I don't think I've actually worked on an enterprise site that didn't have some kind of duplicate content issue. And it's usually things from things like improperly implemented internationalization. Things like help centers or blogs. Even off-site support centers too tend to cannibalize content quite a bit as well. Next up is things like site structure. Maybe you want to make changes to information architecture. You want to introduce new methods to do internal linking. You want to add schema. Fast and Nav is really common in e-commerce sites. While you have performance issues, your sites flow. You have a CDN that's not great. You don't have a CDN. Your site tends to have a lot of surveyors. All performance issues that tend to impact really big sites. Then of course, the issue of creating quality content at scale. And then lastly, maybe backlinks. But that's mostly if you're like entering a new market or you're trying to launch a new product that you don't really have a background in. The thing I want to focus on instead is understanding impact, getting that prioritized, and then getting buy-in and getting things across the line. So when I think about prioritizing the laundry list of SEO initiatives that I could possibly be doing, I always ask myself these three questions. Does this impact crawling or rendering? Does it impact a lot of page types? And is this actually strategically relevant to the business? The first step is actually the limiting step. In the enterprise, that first step is what you should always be thinking about. And the reality is, you know, when you have a small site, Google is not going to have a hard time discovering that content. If you have like a few hundred pages, even a few thousand pages, it's not a really big deal. But when you're talking about millions, tens of millions, 100 million pages, it's really important that you are looking at what Google is able to crawl and making sure that you are optimizing that pack for health. Because the reality is, if Google can't discover your URL, then they can't crawl it. Then they're not going to index it. And then you're not going to rank. So then you're never going to show up. So it doesn't matter if you're like, you know, creating random blog posts. It doesn't matter if you have these like other cool new features in sections of the site. Google's never going to discover those things. You're never going to rank for them. So this can include things like server errors, especially on particular pages. So things like your robots.txt, Google will stop crawling you. If you have certain pages that you really care about, but they keep throwing 500 errors, then Google will actually drop that URL. Overall site health, things like internationalization, fast and nav, or like no indexing sections of the site. And so if you have an issue that's impacting crawling or rendering, then this is always a high impact problem. You should really prioritize this first. If it's not, the next question you can ask yourself is does this impact a lot of pages or page types? Most really big sites have some kind of like templates that the site follows maybe like certain templates for blogs, certain templates for categories, certain templates for home pages. So you have to ask yourself, you know, is this big enough to really matter? So does it affect the IA? Does it affect a lot of page types? Is this meaningfully actually affecting the HTML footprint of your site? If that answer is yes, then it's high impact. And if it's no, then the last question you have to ask yourself is, is this strategically relevant? Now, I just said for it to matter in an enterprise, it usually has to be really big. But sometimes we make special allotments, right? Special allowances for things that are really strategically important to the business. A really good example of this is like in Square, we really cared about like point of sale keywords, Apple Pay keywords, point of sale APIs. In Dropbox, we really cared about cloud storage, right? Basically your fundamental job to be done. These are the things that really matter for us to show up for. If that's the case, then it makes sense. It's high impact. You should prioritize it. And if it's not, then it's probably low impact. And if it's somewhere in the middle where it's like more of like a should do versus a must do, then it's kind of medium, your catch all bucket. A lot of the stuff that we work on is in a gray area. So next up, getting buy in. I think this is pretty hard, but the reality is you only really need to master two simple skills. The first one is forecasting wins. This will help you get resourcing against other teams. And then storytelling with data. This helps build excitement and confidence within your cross functional partners and your leadership. And if you're like me, you hate forecasting, it always feels yucky. But the good thing is you're never going to feel good about it, so just do it anyway. It's not a promise. It's just a prioritization exercise. And the good thing is in a big company, there are pretty much three acceptable ways for you to forecast. The first one is top down. The second is bottom up. And the third is competitive share. So we'll stop with top down forecasting. So top down forecasting, most of you guys have probably been through this. Leadership sets a goal. They say we really want to do this. In this case, maybe like square, they're like, we want to sell 1 million Apple Pay readers. Two, the middle managers then figure out the work and they distribute it to kind of like the ICs. And the pros is it's really easy to get resourcing. This is actually a really good opportunity if you have anything in your backlog to kind of try to get a momentum behind. Maybe you have some site restructuring you want to do. You want to like build a new feature on your blog. This is a really good time to kind of like use this big company momentum to get those projects across the line. The other pro is that it's kind of understood that it's an aspirational goal. People don't tend to like hold the measuring stick after this is done at the end of the year and say like, hey, did we exactly make this goal or not. And the cons is that it's a little fickle. These like, you know, marching orders, company North Star, they tend to change. So, you know, they're a little fickle. Just make sure you're always watching the company's bottom line and you'll be fine. And the other kind of negative is that sometimes SEO might not be the best channel for this goal. For example, when Square wanted to sell 1 million Apple Pay readers, they said, SEO, how are you going to do this? We said, well, there's only 50,000 people that only search for Apple Pay globally. If we sell every single one of those people a reader, it'll still only just be 600,000. Those people are not all looking for a reader. We will not sell that many. So sometimes it's not the perfect channel for the goal, but as long as you're honest, I find that it's fine. Everything works out. Second is bottoms up. This is more traditional SEO forecasting. This is you're saying, okay, we want to own these keywords. You do a little bit of napkin math where you'll rank and you add up the total contribution and present it to your leadership. The pros is that it's usually a little bit more accurate because you're kind of going like keyword group by keyword group. The cons is it only really makes sense if you have a term like point of sale, if you have a term like card reader. It doesn't make sense to do this for your entire content backlog because it's a little myopic. It might not be big enough by itself. Like I said, it only really makes sense for strategically relevant terms. If you want to see what that napkin math looks like, I'm really bad at math, if I'm being honest. My mother was proud about that. But anyways, so you basically look for your primary keywords. You say like, okay, we want to own these keywords. Here are the secondary keywords that support that keyword. This is where we rank today. This is where we rank tomorrow. And then you would build a forecast off of that. And that's pretty much your happy scenario. And you would, to figure out like where you would end up eventually, you would say like, hey, you know, based off of competitors, usually some kind of like off page analysis, you know, you can use h, ref, summer, she doesn't really matter. Based off that, we actually think we should rank here. And the next thing I like doing is giving a high and low band. And the reason why is because it shows that you're being a variable, you know, output might be variable. It's not promised. But you know, different things can happen like seasonality and execution, executional excellence. All those things can kind of impact your final output. But so I really love to give a high and low band. So I just give like a little haircut. Maybe like 20, 30 percent on top of the happy path scenario. And last thing is competitive share. Competitive share is pretty much you're saying like, this is the TAM, the total addressable market. And we want to own x percent by x date. So let's say you want to sell something to doodle dog owners. There's about eight million of them in the U.S. I am one of them. Desk Bailey. She's very cute. She's all over my Twitter. Anyways, so you want to own the doodle dog market in the U.S. There's about eight million of them. You want to capture 7.5 percent by end of year. That's around 650,000 possible users by end of year. And the pros is that's very easy, right? Like, I take a lot of cognitive load to get to that goal. And the good thing too is that executives love goals like this because they love stealing from competitors. It's really easy to tell the story of TAM, right? We have a lot of room to grow. We have eight million people who could possibly use this product. But the cons is, of course, it's a little inaccurate. You don't really know when projects are actually going to hit, but it's really good to get initial kind of excitement about what you're doing. And lastly, if you're still experiencing a little bit of friction, you can propose it as a test. A lot of times, you know, SEOs, we tend to... I think my screen is a little... There we go. Next thing you can do is propose it as a test. I think a lot of times, as SEOs, we feel like an incredible amount of pressure to know everything, to know exactly like what's going to happen, when it's going to happen, when should people see the impact. But the reality is every site is different. Things are executed in a different way. And so, honestly, if we just propose it as a test, particularly with our engineering and product partners, I think you'd be surprised that they're more open to this than you would think. Jen Matthews, who's SEO goddess on Twitter, she... I think used to work at GitHub, is now at another company, had a really great talk with Martin Split, who is the developer advocate at Google. And they were talking about this is part of engineering culture. It's a very normal part of engineering culture to build MVPs because we don't know yet, and we're just going to test and we're going to learn. And together, we can kind of get more projects along the line and hopefully more SEO initiatives funded. And so, last, I'll talk about storytelling with data. I really love this book. It's not an SEO book at all. So, a little background, the girl who wrote this book. I love this book. It's probably been the most impactful book of my career. I buy it for every single person who works for me. If you're at a big company, it's absolutely imperative that you get really good at storytelling with data and making decks. So, Cole Nevek, she was an ex-Google employee and she was so good at making decks at Google that they literally just started paying her to teach everyone else how to make decks. And they would just fly her from location to location. And I really love this book. It's been the most impactful book of my career and it's less than $30 on Amazon, so you should definitely get it. But like quick lessons, using color to draw attention. When you look at this image, where are your eyes drawn? It's pretty obvious. It's the red balloon. It's the only thing that has color. But being really intentional about your use of color. This image is a little harder. Give me a second to look at it. Most of you guys are going to focus on the hot pink sign. It has high contrast. It also uses a really bold typeface. People also, when they don't know where to look, they start reading. They go left to right, left to right, they zigzag down a page. There's other things you learn to, like when people tend to read a page, all they're really looking at is like titles, headers. They're looking at the bullet points. They're looking at what's inside the table. You want to know how to actually get people's attention. So there's small tricks that they teach you how to do there, like with this time series. You can see like, okay, we got rid of the border. We get rid of the grid lines. We soften the y and x axis. It makes it a little bit easier to see the trend. You have other options too. This is just like a spaghetti graph. Maybe you want to use color to just draw attention to one of those lines, something that's happening. Or maybe you just want everyone to really focus on like what was out, but what happened at the end of July in this case with monthly sales. Or if an executive asks you a question or you're trying to report on progress. For example, like this has definitely happened before, executives ask, you know, why does SEM contribute more revenue to like this particular product than SEM? If you just do like a basic export of you know, all the different kind of keyword types that SEM is bidding on, you know, what will never happen when a lot of you guys are probably doing is like scanning the pie chart right now. I mean pie charts are not great. You know, most of these things you can do in a bar graph, but we're just going to do a pie chart, for example. But if I just change this a little bit and I gray out the areas that I don't really want people to focus on, I want to say like, okay, 14 percent of those keywords, we actually can't legally say them. 25 percent of those keywords are competitor brand keywords. We actually don't qualify for 39 percent of those keywords, so of course we contribute a little less. Makes it a lot easier to focus their attention on the things that the message that you're actually trying to deliver. And lastly, again, if you're having some issue getting buy-in, think really critically about the initiative you're trying to push across. A lot of times when we're doing SEO planning or we're doing our SEO road maps, a lot of times we end up putting things that are related to site health in those road maps. But the reality is engineers care about that too. Web engineers would monitor site health, even if it wasn't for us. Tools like Kibana or Grafana, all these kind of like logging tools, observability tools, exist even with it outside the SEO world, and in fact they're a lot better when they're not built for SEOs. And this is things like excessive server errors. Google can't access your content. You are blocking good bots that need to be able to crawl the site. Those are actually all bugs. They're not SEO initiatives. So if you have any of these issues, then it might make sense to actually go to your engineering team. Go look at their JIRA and file this as a bug, because that's a bug. That's not an SEO initiative. And prioritization. So lastly now you have like your marketing leadership really excited about all the SEO stuff you're going to do because you've did a really good forecast. I mean you told a really good story. Now you need to get engineering and product bought in. And the reality is most of the time product engineering lives in a completely separate org as you. So how I like to prioritize with my engineering and product team is a little bit different. First I want to understand their goals. That's because sometimes they have a different goal than us. So it's nice to kind of start off with what does success look like for both of our teams. Number two, I asked for an open idea submission, collect ideas. One of my favorite ways to brainstorm just because I generally hate active brainstorming like sitting in a room getting everyone to brainstorm is we will get everyone into a room, a quick 10 minute meeting, and talk about the problems that we want to solve or talk about our goals. And then we have everyone come back a week later and submit all their goals into a sheet. This gets everyone a little bit time to have creative ideas. Some of us you know are not really good at thinking amongst other people. They will need like a little bit of shower time, whatever it is. I really also surprised too by the submissions sometimes from like the engineering team or submissions that we get from really junior members of the team. Sometimes they can contribute really interesting ideas. And maybe you're not going to take that idea, but maybe you guys can work with it and play with it to make it like a really fun impactful project. And then lastly, use rice to prioritize. Really popular with an engineering and product orgs. So first off, what is rice other than a delicious carbohydrate? So rice is a product priorities prioritization framework that's very common within product orgs. It's used globally. It's not just something that's based in the US. So it's very familiar. It stands for reach, impact, confidence, and effort. If you want, you can make it a little bit more simple and you can use ice instead. Impact, confidence, effort. So once you have all your ideas, I like to put them into a G sheet. And what you do is you score this particular initiative against these four different dimensions. What you want is to make something good be one. I like using one and making the good thing being one, a smaller number just because you could do it the opposite way. You can say five is good and try to prioritize the biggest number, but it's a little bit weird when you talk about effort because then effort will be a big number for good. It's just very confusing because a big number for effort seems like it's bad. So I like making it the number small. So in this case, you want to do something that affects the IA. That has a big reach. We're going to give it a one. We think it's going to be impactful, also one. Confidence, very confident. We're giving it a one. Effort, very challenging, very, very challenging, especially when you think about something like internationalization, you have to translate all those URLs. It could be a really big effort. You have to make it dynamic if you're moving things around as you do in the enterprise. So it's overall going to have a score of eight. Then you basically stack rank all of your initiatives based off of the score. And at the end, what you'll do is you'll sort so that the smallest numbers float to the top. These are the most impactful, the highest confident projects. And I find that this is really popular with all my product engineering partners. I also like that it's fast. I can do this literally in under an hour. I mean, we'll have our whole H1 or H2 plant. It also makes it really easy to know what's in your backlog. Like, let's say you finish things early, resourcing opens up. You automatically already have a list in your backlog. You'll know immediately what you already decided was really important. And this actually usually works out so well for us that we very, very rarely go back and reprioritize projects. It's also really easy to share an update. And it's really transparent. So when someone comes to you and says, like, hey, we want you to work on something else instead, you can say, like, hey, where are you in this stack rank? It makes it really transparent for your cross-functional partners what you're working on. And again, PMs recognize it and they understand it. And once you're done, I think this is a really important part, which is share the wins. I really love this. I think that it's an important part of any SDO's job within a big company to follow up with their cross-functional partners. And that's following up with your engineering team, following up with your product team, and just letting them know, first off, thank you. This was the impact. There's actually a study that was done where if people are even just thanked, if you just thank them, 66% were more likely to respond positively to a second request, versus only 32% were willing to respond when they were not thanked. So I think thanking is definitely underappreciated. So I'm trying to make sure everyone's thanked. And if we have a win, I try to share it with everyone. I try to let them know we have a win. I CC their boss, so that their boss can know that they had a win too. And it's awesome. I've always find that it doesn't dilute the nature of the win. A lot of times it's, you get bigger and funner wins together. And everyone ultimately will know that, you know, it's an SDO project. So don't worry about it. They're not taking anything from you. You guys are just building a better relationship. It's okay, cool. What does this kind of look like end-to-end in a really big company? I think I have to start going a little faster. But at Dropbox, a little bit of context we wanted to build an internal linking tool mainly just for the body of the page. So not structural elements because we had already linked in the structural elements. I was supporting a major product. So this product was already in the breadcrumbs. It was already in the IA. It was in the Flutter. What we wanted to do was actually hyperlink words in text in the body, but they kept getting written over by authors, of course, who didn't know why we had the ugly links there in the first place. And so we wanted to create an internal linking tool. So the first thing we did was we manually linked. So that included things like product pages, blog articles, some help center. We did just a site colon search inside Google to find where on the site do we say this word, and then we hyperlinked it. And while this is easy, it's not super scalable, but it gave us kind of our first win, screenshot. This got the product org a little excited. So now they wanted to help us build a linking widget. So if you look at the product pages on most Dropbox pages, they'll have a cute little widget where it looks pretty much like a related articles module. And we could link to other product pages to each other. This was really great, but one of the cons was it was really manual. Still, we had to upload the image. We had to choose the anchor. So while it was really good for the diversity of anchor text, it was still really manual. The other great thing, though, about this widget that I want to highlight is that it also really helps conversion rate as well on these pages by not pigeonholing people to this page. So we shared all those wins. We shared even the wins that were not SEO-related, strictly SEO-related, like the conversion rate win. So we shared. We shared it in emails. We shared it in all hands. We shared it in newsletters. It's really important. If you tell something, if you tell people in a big company something once, they'll almost certainly forget it. And then finally, we got the MVP for internal linking. So a lot of times when we build things in engineering or a lot of times we'll build an MVP that doesn't have quite the full functionality of what you envision long-term. But basically, we had a CSV and a system in which to upload it. A lot of times, these tools are initially built with using a CSV or an Excel file just because that way the engineering team doesn't have to invest in front-end resourcing or a UI. And we would tell them, okay, take this word, hyperlink it to this page. And it was a pretty foolproof way of actually getting links on the page. And it allowed us for a lot of diversification of anchor attacks. The cons were, of course, you know, if those links ever had to change then we would need to manually update them. And of course, it wouldn't be able to handle translated URLs by itself, but we saw a really good impact. We saw anywhere from about 15 to 60% increase in organic sessions and entrances for targeted pages. Pretty much every single page saw some positive impact. So if you want to build something like this, it's actually pretty simple. And it was great because, like I said, it made it so that just the regular SEO manager could go into the CMS and automatically hyperlink this word across all of our different page types. And it could be well-sprinkled throughout the body of the page. So if you're going to do something like this, just make sure that you're not overwriting existing links. You're adding variants and synonyms to that page as well. And just know that you only really need to link one link per page. Don't go crazy with it. There's not a lot of incremental value to doing it, probably more than once in bodywork makes sense. So when you are on the Dropbox site, you'll actually see a lot of those links that are happening in text are actually just fully automated by this very simple CSV upload. So again, just rinse and repeat these three steps. Identify the best SEO opportunity. Make sure you ruthlessly prioritize. Nothing engineers dislike worse than getting a laundry list of different errors that like screaming a frog or something found on your site. Then you want to forecast your potential wins, storytell and get buy-in. And then make sure you share the wins and screenshot everything. I always regret not having enough screenshots to tell my story. And last, we'll talk about doing Enterprise SEO in a post COVID world. Of course, Enterprise SEO has not really changed during COVID, of course, but our lives have. It's certainly affected my life personally and professionally. And a lot of what I'm trying to do these days, I feel like is just take back some of my time. And so before I leave, I want to share two ways that I'm kind of doing that. So the first one's using a tool called SEO Writer. I am obsessed with this tool. Basically what you do is you upload a list of URLs. I do one URL per country, per page type, per language. And then you basically just send it to the system and you can tell them how often you want them to check this page. You can have them check it every day. You can have them check it once a week if you want it to be a little cheaper. It's actually very affordable. I do it Monday to Friday, so that way we don't miss anything on the weekends. And it looks for SEO for major SEO elements on the page, things like metadata, meta robots, robots.txt, schema, hreflang. And it looks for all these things for you. It even looks for things like did your page take a lot longer to load? Did your word count change? Does somebody just inject a bunch of words onto the page? Or did they remove some that might be important? And then you get these alerts. You can actually click into the alerts directly from the emails and it shows you like, oh look, the title tag was this value on this day and now it's this. I love it. They also have some other cool tools, things like fetched versus rendered. They can touch your staging versus live site. A lot of really cool tools in this. But one of my favorite things is the different ways you can deliver these alerts to yourself. So of course they have email alerts, but they also have Slack integration so that you can actually send it to your team Slack channel so you guys can triage these actively and you know if someone is already on it. For really high critical alerts, things like new rules, the robots.txt were added. A page was no indexed. We actually get these texted to us. All the leads get taxed saying like, hey, somebody no indexed a page on your site. And so I love it. It makes it really easy to monitor a bunch of sites or even your competitors if you want. Not like I would do that. Another thing is the average enterprise SEO also spends a lot of time in meetings. Of course you have a lot of cross-functional partners. You have a lot of stakeholders. I did a quick survey asking SEOs how much time they spend in meetings. So I was feeling particularly bogged down that week. And about 34% of you said you spent 16 hours a week in meetings. 21% said 15 hours up to 15 hours. So a lot of time in meetings. One way we've been kind of reducing this at Uber is I'm really lucky. I got to work with a lot of really smart people every day. And what that means is we can move most of our readings to more like a bi-weekly or tri-weekly cadence. So for weeks that we don't meet, I actually instead have these Polly stand-ups. So Polly is a free app on Slack. There's like a bunch of different stand-up apps that you could possibly use. And it really asks just four simple questions. And this is pretty much replaced most meetings for me with senior people. The first question is what do we complete? Second question is what is our focus this week? Third, are there any blockers so that I can like escalate them up and down? And last, are we on time to meet this goal? It's usually some kind of milestone, some date. And this is really all I really ever needed from that 30-minute meeting plus some prep anyway. And it's really helped me take back a little bit of my time so I could keep doing SEO. That's it. Thanks, everyone.