Several months ago, we brought in our SEO team to answer your questions around developing an SEO strategy and ensuring your site stays at the top of key search results pages. And now more than ever before, as organizations move even more of their business processes online, itâs crucial to get found in Google.
On Thursday, Victor Pan returns to the Power User Community to answer all of your questions on SEO. At HubSpot, he helps us rank at the top of thousands of search results pages. He and the team align our content strategy with search demand, build new tech infrastructure for HubSpotâs web properties, and help to grow HubSpotâs SERP presence internationally.
Start asking your SEO questions now, and Victor will answer your questions on Thursday!
I appreciate you may not be able to answer my question but I thought Iâd put it out there anyway for discussion..
When Google Pagespeed Insights started to use Lighthouse as an engine to run their performance audits, our scores dropped from 83/100 to 39/100 and it was pretty concerning, but then I ran Hubspotâs homepage through it and it got an equally poor score. Now however, the Hubspot homepage scores about 94/100 and weâre still on 39/100. So I was wondering what changes Hubspot had made to their site to improve this performance and were any of those changes being rolled out to customers who host their site on Hubspot COS.
A particular point that comes up a lot on SEO audits for us, is Minifying Javascript. Looking on the page editor and header of the site, there werenât any attached javascript files that we had added, so it doesnât appear to be something we have control over. Rather just how Hubspot loads the site in the backend. Itâs sometimes difficult to know which audit points we can fix ourselves and which cannot be fixed due to the code being on the backend of Hubspot COS.
Another question would be about webp files and whether Hubspot is supporting them or if there will be a converter built in to optimise from jpeg to webp for example?
And another question around AMP. Is there any information about which Hubspot elements donât work on AMP pages, for example pop-up forms, livechat etc
Excellent questions! LBeresford-Ward
Iâm going to split this to three answers into separate replies. I joined HubSpot around 3 years ago and I have experiences working in fast moving startups and slower moving enterprises. HubSpot was inbetween the two company phases when I joined.
âWhat changes Hubspot had made to their site to improve this performance and were any of those changes being rolled out to customers who host their site on Hubspot COS?â
TLDR; HubSpot improved PagedSpeed performance from the top-down by working with web developers to minimize request files and sizes, product engineers to address render-blocking javascript which is a part of HubSpot CMS, and content creators on image SEO best practices.
PageSpeed - HubSpot is one of HubSpotâs biggest customers. PageSpeed Insights is a fickle tool unless you use the same connection, device, and incognito to run it. Definitions within PageSpeed Insights have also been evolving, so thatâs been fun communicating the need to benchmark consistently. Itâs similar to the question the SEO-team used to get on âwhy isnât HubSpot showing in position 1 for Inbound Marketing?â -
There is a task force focused on performance (Iâm semi-involved) because you canât âfixâ performance by yourself. Itâs an evolving concept as new technologies are added to websites to meet customer needs, and multiple stakeholders can influence performance.
Render-blocking resources (really jQuery) was the big one that dinged HubSpot (and HubSpot customers) along with having over 200 resources being loaded, lots of sub-optimal images, and a huge page file size. The impetus for this change across the organization was showing a video recording of our homepage load on a 3G mobile device via webpagetest.org, and then we got working.
Our product teams worked on providing our customers the option to move (and update) jQuery: Include jQuery across pages - They also worked on automatically combining and minifying CSS/JSS using brotli compression (better than GZIP performance-wise) via our CDN. Thereâs a fuller list here: Site Speed and Performance: What you can do, and How HubSpot Helps and a lot of it is in the backend.
Based on your description, I would check to see if youâve inlined CSS/JS or you use 3rd party JS that gets added all over different parts of your websiteâs templates. We had quite a few of those within HubSpot.
Our web development team (for marketing) worked on getting those 200 requests downs and tested the net impact of lazy-loading images. Theyâre the true heroes of most of the changes in that score.
Our content & SEO teams worked on creating guides on how to pick, choose, resize images and we scoured all legacy CTAâs, posts with gifs, etc. and looked to clean up the old but make sure net new content would not be too expensive for the rest of the world where internet access is more expensive. We have process documents for training writers who are goaled traffic generation.
Our analytics team now holds the keys to adding 3rd party tools onto pages (and logins) - so that itâs all centralized. We traded a bit of âmove fastâ for some centralization and governance that saves us the tech debt in the future.
The work here isnât done, because weâre awefully aware that our customers look up to us for inspiration. Note that you might have to dig around on some of the settings. Usually we err on the side of donât touch our customersâ settings.
I also struggle with PageSpeed as well but unlike other large companies, I am the sole SEO employee & the only person on our marketing team who has any dev experience at all. Iâve done as much as possible to optimize content & hyper-focus on E.A.T for an industry whose keywords are often too low for Google to even monetize.
Iâve added Structured Data & Schema everywhere possible, but my department manages 3 main brands + 2 - 4 ancillary brands & the site templates that our core brands were originally built on were not overly mobile optimized so itâs a constant struggle between good UX & good Dev practices.
I guess my biggest question is, in a company like mine where SEO is not recognized as a âneedâ and where 3 marketing team members (Director, Graphic Designer, Digital Marketing Specialist) support 32 locations, for 6 brands currently in 12 states⌠how can I convince our executive team that SEO is vital?
Thatâs a great question on stakeholder management. So one of the things I started working on is reasoning by first principles. In SEO, the stakeholders always end up being searchers, the search engine, and business owner and you have to prioritize between the three needs.
To get executive buy-in within HubSpot we always prioritize the searchers - our would-be customers. Iâm privileged to work in an environment that doesnât put profit over people.
At an agency serving P&G, it was about both doing right for the brand and shining a light on the Brand Managers we reported to. This meant the brand managers had appetites to do something new and innovative and their needs were proritized over searchers or search engines.
I would recommend learning what your executive team is held accountable for. For example, if itâs P&L, then creating a reporting system that shows the ROI of SEO would create the change you need. You donât have to start with fixing the whole thing - start with a topic cluster. X traffic lead to Y signups or Z downloads/activities tied to $$$. If you donât have this data available, SEOâs typically use hypothetical PPC spend to get the same amount of traffic to show the recurring value of SEO.
Once you have that established, you can start talking about all the activities preventing you from increasing that number. This will lead to either process improvements, team budget increases, or more headcount.
Money is often the easiest thing to talk about at the executive level.
âWill Hubspot is supporting webp, if so, will there be a converter built in to optimise from jpeg to webp?â
So from what I know, you can throw webp files into file manager. In terms of asking for performance boosts by detecting browser client (Safari is still not on the webp train according to caniuse.com) thatâd be something to add into our ideas forum.
I think itâs a great idea. I have no idea what thatâd cost for hubspot, but Iâd upvote that. I think our engineers would love that too, because itâd just work like magic for our customers.
On the search side of things, it gets a bit more complicated. Googlebot will now crawl as evergreen Chrome. If we only ever serve webp to search engines, Iâd have to test this, but I canât imagine a webp search result within Google Image Search being a good one for someone on Safari (Apple does not support webp). You bet Iâll get involved with our product team if we decide to support webp based on the browser client.
âWhich HubSpot elements donât work on AMP pages?â
Iâm going to narrow this down into the AMP variant of blog pages described here: Use Accelerated Mobile Pages (AMP) in a blog
Iâve been advocating for editable templates for AMP within HubSpot, and ideally beyond just the HubSpot blog. AMP as a format is extremely restrictive on javascript. A few years ago forms werenât even supported. AMP is an emerging technology where adoption is heavily skewed towards news media websites, and there were some rich result-specific search results that required AMP that no longer require it.
The ideal scenario would be that all those things you love about HubSpot which require javascript would have AMP equivalents. The reality is that we have to prioritize - and the easiest way to get that prioritized as a customer at HubSpot is to get upvotes in the ideas forum.
HubSpot is not an AMP-first platform. Weâve simply created a mechanism that strips existing content into an AMP format. I believe weâll go beyond that if thatâs where the industry is moving towards or if enough of our customers request it. Thatâs been my experience working with HubSpot product organization. Yes, solving for the customer is more importing than solving for enterprise value, but solving for some customers over others is a tough balancing act.
Get out there and vote on some ideas! You paid for HubSpot, now go make your vote count! We can only be as data-driven as much as our customers are willing to be transparent to us as we are to them.
@Pagenoi , @Rolf , @lindseygarrett based on your intros, it seems like you spend some, if not most your days focusing on SEO. Any questions for HubSpotâs SEO expert, Victor?!
Really quick question on HubSpot practices regarding blogs and landing pages. There is significant evidence from different SEO communities that consolidating all content into a unified TLD really helps improve SEO rankings. When creating landing pages and blogs in HubSpot, with a website hosted elsewhere, you are required to use a subdomain configuration which the majority of SEO communities recommend against doing. What is HubSpot rational behind this decision? To add an additional layer of complexity, regions like China require an ICP license to advertise on Baidu and are given preference in SEO rankings. These ICP certificates only resolve to the TLD (www.domain.com) and do not cover subdomains (blog.domain.com). Your thoughts.
Hey @asterix2x ,
I hear you. As someone who has lived and worked in Shanghai most of the western world arenât familiar with how fragmented the internet is. Thereâs no easy answer to that specific scenario, especially with the great firewall and censorship requirements to do business, let alone some of the cultural practices that are would be considered bribery or quid pro quo.
So the technical reason (security and HubSpotâs infrastructure) is here: https://community.hubspot.com/t5/APIs-Integrations/Why-Can-t-CMS-COS-Pages-Be-Hosted-On-Root-Domain/m-p/227257
TLDR; Weâre working on a solution, but there are security reasons (DDoS) for how HubSpot is set up to manage and keep our customerâs domains secure.
Frankly I have little to do with that decision because itâs before my time and I do not have deep expertise in that area.
What I do have a good track record on is ranking on the same term on the first page with multiple websites, consolidating websites into a single TLD, and keeping organic traffic at 100% pre and post migration. My experience shows both ways can work, and so does the experience of the team at dotdash that spun About.com into several properties: https://blog.searchmetrics.com/us/about-com-domain-splitting/ - itâs important to note that what could be true 5 or 10 years ago (one TLD versus many cctlds or subdomains) isnât true today. The debate never ends because the timelines are evolving.
Regarding SEO, to those who saw improved rankings from consolidating content into a unified TLD, Iâd like to note that itâs almost never a 1-1 comparison. Theyâll consolidate and also 301 redirect a blanket of 404âs along the way, redesign something, or update broken links within redesigned navigational elements. This is true about the dot dash example I referenced as well but it does show the URL setup isnât as big of a SEO detriment as people put it out to be. If anything, you can have an edge given that there was an update to the domain clustering algorithm of 2012-2013. Theyâre calling it the domain diversity update (2019). You can rank for the same search term with two sections of your website targeting different intents.
Searcher intent is often fractured for head terms.
Victor,
Iâve asked this question too pretty much every expert I can find and most of the advice is similar but Iâm curious if from an SEO standpoint you have any additional insight.
Industry Details: I work for an HOA/Condo Association Management Company, we contract with HOA boards to help them manage day to day operations (inspections, financial, SAAS solutions & various other duties). We are technically âProperty Managementâ but thatâs a word used in our industry for leading management & generally isnât an effective keyword for us. Our highest non-branded trafficked Keyword is âHOA Managementâ - industry-wide & good considers it pretty low quality (2,400 monthly searches maximum).
Company Background: We have 3 core brands, our main management services brand (RM), our âwhite-gloveâ services brand (GM) which serves the same market but is like our âluxuryâ division, and our SAAS (CC) - plus we have 3 ancillary brands that will eventually be absorbed into our main brand (RM).
Issues & Questions:
1. One of our biggest hurdles is Content Cannibalization, specifically between our two services brands (RM & GM) where the only demographic differences in our clients is generally our Luxury division (GM) have more units & larger budgets for amenities like on-site concierge services and there are significantly fewer of those clients. We publish a ton of blog content and have optimized our content strategy for featured snippets & other SERP features. - The issue we run into fairly often is overlapping content that creates competition for ourselves. We use âCanonicalâ links as often as possible but that always forces us to choose which brand will get the actual link equity. Have you ever dealt with two brands that have the same basic keywords (so few long-tail are even acknowledged by Google itâs difficult to differentiate) and share a demographic? What kind of things can we do to keep from competing with ourselves.
2. There are very few high content aggregators in our industry, and even fewer with good DA (the average is about 35 & highest maybe 55). What could we potentially do to increase our backlinks when the community is so small?
3. As I mentioned earlier our highest trafficked Keyword is âHOA Management Companiesâ with 2,400 (optimistic) searches per month, long-tail keywords are often too âlow qualityâ for Google to even rank, though we still optimize for them. I can increase traffic by including Property Management keywords but even if we got more traffic itâs highly unlikely any would ever convert, but are there other ways we should be approaching keywords?
4. One major backlinks issue we have is because of our âPortalsâ - Many communities have websites & they link to our Portals, but as this is part of the SAAS, they link directly to the database (portal) login page, which means 1,000âs of quality backlinks go to our portal login page which is not our main domain. In my opinion this is a waste of link equity, as the portal pages arenât even really indexed. My team has denied me the ability to insert any kind of integrated login, including Iframes, based on âsecurity concernsâ. My last ditch effort was to create a URL path on our brand domain that just forwards to the login page, but even then the current 1,000âs of current backlinks are unlikely to change. Any suggestions on persuading them that itâs important?
5. Finally & this is just a more opinion question but Iâd love to know anyoneâs thoughts on it. For a company with 3 Core Brands (3 Hubspot sites), 32 branch locations spanning 12 states all supported by a central corporate marketing department, how many people do you think youâd need to maintain good SEO? How many hours do you think a company with those specs should devote to SEO/Content Marketing.
Thanks for any insight you or any of the other members here can provide!
First of all, @Pagenoi you sure wear a lot of hats! I have a feeling youâve already got an answer you want to hear and Iâm not sure Iâll be providing that.
Problem 1: Content Cannibalization
TLDR; Keep on with your syndication strategy using canonical links. Alternatively you can consolidate with redirects if canonical links are being ignored by search engines provided it matches the search strategy you have planned.
Content cannibalization isnât an issue when youâre in positions 1 & 2 - itâs an issue when youâve got multiple pages that donât even breach page 1 of search results, or have to weigh the potential gains of consolidating two results on page one in hopes that the net result is an even higher ranking. You should test this and stick to your learnings over industry ponderings.
Iâve dealt with 7 brands with the same keyword set. We intentionally had 7 different paid search strategies for bidding on the same keywords as well. We were not very ROI driven - we had clients that wanted an 11X increase in impressions for the brand within a year, so we delievered.
Thereâs a marketing adage where you have to see an ad 7 times for it to leave an impression on you, they equivocated impressions in organic search and we delivered.
Then the next year there was the decision to simplify branding so we migrated them all into two places. Those websites are crest.com and oralb.com. Crest Pro Health, Crest Complete, Crest 3D White, Crest, and probably something else that escapes my memory since they also had WAP-friendly em-dot websites. Each of those sub-brands worked with different personas and had different editorial voices - so imagine 7 articles about cavities or gingivitis. It wasnât seen as competing against ourselves - it was seen as crowding out the search results. To be the number one most trusted brand meant youâd have to be the number one most common brand on a Google search. We expanded that to also include branded mentions on sites we didnât own.
Problem 2: Higher quality backlinks?
Answer: Get on the news. Thereâs a domino effect of news syndication and folks commenting on the news. If I can get quoted on Mashables with a backlink to the company I used to work at with a bit of creativity, then so can you. Youâre much more skilled than I was 7 years ago.
If you struggle with creating linkable ideas, then hire someone who knows how to do it right. You already have a lot on your plate. Maybe itâs not about getting more backlinks but getting more out of what you already have.
At some point we ran a test at HubSpot where we added internal links from our more popular blog authorâs bios to their âfavoriteâ HubSpot product and we saw ranking improvements on competitive terms like âcrm software.â Make do with what you do have.
Problem 3:
Iâd start with doing a total addressable market analysis. Then you get through a phase that gets a bit philosophical. Lifecycles. e.g. you figure out the life of a query before and after theyâve reached your website and expand that way. Start with Google Search Console. Expand with People Also Asked (PAA) for every single query that you have in GSC. Look for overlaps in PAAâs in that query set. Test out what converts and find patterns.
At HubSpot our free tools are a great example of lateral forms of aquisition. Website Grader, Invoice Template Generator, and Email Signature Generator come into mind. Content can exist in many formats.
4. An SEO at HubSpot agrees with @Pagenoi that thatâs a lost SEO opportunity. Iâd defer to that author bio experiment I mentioned - run a similar one where you have a piece of content that is heavily internally linked or has a lot of backlinks - and you add an internal link from that page to another page - and show ranking changes. This would be similar. Now you just have to work the people portion of things so that theyâd care about it. Having a tangible $$$ tied to it always helps. Note that this is a change thatâs purely for SEO and business owner and not the searcher.
5. Tough question that depends on the acquisition strategy and tech stack. When I started HubSpot had around 2,400 employees and one person with SEO in their title (me). Iâve lost count of how many people work at HubSpot now but thereâs now 8 folks with SEO in their title within HubSpot. We canât go any leaner because Iâm not a polyglot that also speaks French, Spanish, Portuguese, and Japanese. I canât liase between product, content creators (blog, academy, support, developers, etc.) and tool creators effectively as a single person. I have and will drop things that need to get done because itâs not being measured.
So now I just focus on organic acqusition tactics that donât require additional headcount to scale linearly. We still need all those other channels and employees directly responsible for the numbers in them, but HubSpot isnât set up so you can be a one-man band. It sounds like youâre in that situation as well where you want to scale-up, but the strategies you have in place requires additional headcount for linear growth. A good example where linear growth is true for HubSpot is the HubSpot blogs. A good example where non-linear growth happens is the simple www.hubspot.com/resource Resource Library we hacked together to make sure no offers get orphaned on HubSpot and old offers eventually get updated/redirected.
So I know it doesnât answer your question, but what I did do was map out what I thought the SEO Team would look like should we hit a few milestones and what numbers weâd be ready to commit to growing if weâre given the headcount. It goes back to $$$'s
Thank you so much for the detailed response and all of the examples.
Everyone on our team wears a lot of hats but I enjoy the SEO and Analytics side of my job very much and itâs definitely what I try to focus on, as I personally see it as one of the best ways to showcase the true ROI impact of good UX and coding.
I honestly had not really thought about the solution to content cannibalization as being to crowd out the SERP with our brands. Itâs a unique industry in that our real competition is fairly small so being able to produce more content and focus on being the authoritative voice in the industry is definitely something I see as achievable.
I am definitely going to be working with my director on the ideas for creating tangible ROI reporting that will be more effective at getting our executive team on board. Again thank you so much for your invaluable insight and the amazing ideas youâve presented!
I have not managed SEO in about 4 years, but am now working on an initiative to transfer our blog to Hubspot and I would like to make sure weâre optimizing our blog posts for SEO using the most up-to-date guidelines. Two questions:
1 - What are your top 3 tips for optimizing a blog article for SEO/what is the most important?
2 - What keyword/planning tools do you recommend most for identifying blog topics that have an SEO opportunity?
Hi @baylistefl ! Great questions,
Top 3 things optimizations to do with blogs.
1. Historical optimization.
2. Periodically prune outdated content that have backlinks, but drive no or little traffic. If thereâs no close variant then youâd redirect it to the pillar page per HubSpot parlance.
3. Have an editorial calendar, content inventory, and total addressable market all mapped out for the year. Think about who will want to link to you based on what you will write about. If you have the budget for a rank tracker, great - if not - leverage the HubSpot SEO tool along with its integration with GSC. For HubSpot, weâve created a strategy for Featured Snippets and rely on 3rd party rank trackers to check on Featured Snippet health (wins/losses). All People Also Asked are Featured Snippets, but not all PAAâs have significant search volume. At the end of the day traffic to your website is more reliable than search volumes from 3rd party tools.
Keyword Planning Tools
HubSpotâs SEO tools uses SEMRushâs data: Validation metrics in the SEO tool. Internally withing HubSpot marketing we use both Ahrefs and SEMRush - the important thing to note is we donât rely on their proprietary metrics. What theyâre great at is aggregating a lot of data. The most important thing you can do after you find those âhigh volumeâ terms is to look at the search result and empathize with the searcher. Their groupings will often not match your business.
I would hire a human being to be in charge of the actual keyword research (donât outsource keyword research) because theyâll understand the business and whether thereâs potential intent from a search term that could translate to a business lead/conversion. The best opportunities arise when the tools are wrong but youâre correct - and your competitors will miss them.
*Edited some spelling.
A company hosts their website on WordPress and Blog on Hubspot. Right now the blog lives in the ânewsâ section of the website navigation bar and Iâd like it to be present on the navigation bar separately as âblogâ.
2. Now if we move the blog to a new subdomain - www.blog.<companyname>.com/.. it would hurt the SEO, right? How can we separate the blog from news without hurting the SEO?
3. The blog has been active since Oct 2018 and has around 80 posts. There is a lot of sessions and it wouldnât be fair to lose all that data.
Please advise.
Hi @AM8
Sounds like a fun problem. So Iâve done migrations with 0 traffic loss before, many ways. Iâve also done migrations where I saw a 20% decline, followed by full recover in 3 months. Iâve also done migrations that fell 80% against my advice. It feels great when you put in the time and it just executes perfectly. It feels painful when things go south, especially when youâre set up for failure. The difference between the ones that kept traffic consistent were:
1. There were 1-1 mappings. If you donât migrate certain URLs via 301âs youâll just lose that traffic. For example, everything to a âhome pageâ is a migration, but itâs a horrible one.
2. The content were of close equivalents in 1-1 mappings. Donât call removing half the copy for certain pages a 1-1 migration. Itâs not.
3. Internal links were updated. Namely in the navigation - sometimes they werenât. Sometimes relative paths broke and created internal 404âs.
4. The sites were relatively healthy in technical health - limited internal redirects, 404âs and mainly just 200 status codes within internal links.
So with that in mind, you will need three things:
1. A comprehensive look at what links to where on your website - both internal and external. At the time of migration you want to update any old mentions to the new URL, and 301 any inbound links from the old address to the new URL. This is true for canonicals as well.
2. Reporting on traffic for both the old site and new - setting this up in Google Search Console is a great idea. Googleâs guidance on change of address is good. You should be seeing two lines basically invert from one to the other.
3. Stakeholder alignment. Youâve got a document that maps out all existing URLs, all net new URLs, all URL changes that have to happen for internal links, all the 301 redirects that have to happen, the sitemap changes, any robots.txt rules that might need to be updated, any canonicals that need to be updated⌠you track the completion of these tasks, the reporting is shared with the full team, and youâve started notifying search engines (Bing has a site move tool too) of the change. Ready? Set? Go!
Now get ready for anything that could go wrong and what would the follow-up be if something does go wrong! Youâve got this.
SEO Migration Nightmare Story
I once worked on a project where Accenture won the website migration contract. I did the strategy and planning, they did the execution. They then outsourced the contract. Iâm not sure how many levels deep this goes, but everything that was broken down into actionable steps werenât being implemented at the day of our migration launch.
Several phone calls later I was talking to a guy with chickens clucking in the background and louder-than-megaphone plastic keyboards clacking away in the background as we called to talk about why the list of 10K+ list of URL redirect mappings werenât implemented. Did I mention there was a rooster? The developers on the other side insisted everything had been done.
We go through the first line. Itâs working.
We go through the second line. Itâs working.
I ask on the call to jump to the end of the list. Itâs not done.
I ask to jump to a random line item. Itâs not done.
They agree to make sure itâll be done by the end of day.
Moral of the story? Be sure to have processes in place to check that every task is done to spec! That and it was absurd how consultancies would win big contracts but subcontract out their work - and some of those subcontractors would subcontract their work. Yikes!
@VictorPan
Thank you for such a detailed response. It was very interesting to read through the Accenture story.
The move seems a little out of scope for me to pull off. I guess the blog will continue to have to live on news.<companyname>.com/blog/âŚ
Thank you for taking the time to answer my question.
Thanks for the opportunity to ask some questions.
Weâre relatively new to the SEO game having some basic understanding of SEO.
Weâve just taken over from external company who was managing our online presence and brought it in-house to create a more tailored and less generic content plan.
Bit of background:
We currently have an externally hosted website and we are using HubSpot Landing Pages, Blog and CTAâs to create an integrated feel for our online content. We have just hired a content manager who will be looking to develop our blog and other online content.
We have been performing some site audits on Ahrefs and have now started reviewing content but I wondered if you wouldnât mind answering:
What should be the first steps to approaching an SEO strategy plan?
Are there any common suggestions that you actually believe is not good or perhaps provide very little gain? There is a ton of online resource but sometimes it is hard to grasp the value in each suggestion
Do you have any recommendation for good resources on managing and developing a good SEO strategy?