How Content Audits Help The HubSpot Blog Age Backwards — A Peek Into Our Process

Trending 1 week ago

In 2023, my squad and I began moving connected possibly 1 of nan astir eager contented audits ever conducted connected nan HubSpot Blog. We’ve tally content audits successful nan past — but not for illustration this.

content audit connected nan HubSpot Blog represented by a magnifying solid and a manus drafting from constituent B to constituent A

We ran nan audit successful 3 phases:

  • Phase 1 addressed our oldest content.
  • Phase 2 evaluated our lowest-performing content.
  • Phase 3 assessed nan worth of our taxable clusters.

When it was each said and done, we audited complete 10,000 blog station URLs and complete 450 taxable clusters.

 How to Run a Technical SEO Audit

In this post, I’m going to attraction connected shape 1 of our audit. I’ll locomotion you done really we audited our oldest contented and really we took action. Plus, I’ll stock nan results we found.

But first, fto maine springiness you immoderate inheritance connected why we decided to tally an audit of this magnitude.

Why We Audited

It each started successful early 2023. At nan time, my squad was called nan Historical Optimization squad and we sat astatine nan intersection of HubSpot’s SEO and Blog teams.

We were responsible for updating and optimizing our existing blog posts and uncovering maturation opportunities wrong our library. (We’ve since evolved into what is now nan EN Blog Strategy team.)

In lawsuit you’re caller here, nan HubSpot Blog is HUGE.

For context, nan blog was location to 13,822 pages successful February of 2023, nan period we began our audit. Ahrefs’ Mateusz Makosiewicz moreover declared it arsenic nan “biggest firm blog … ever” successful an SEO lawsuit study earlier this year.

9 Surprising Takeaways From Analyzing HubSpot’s SEO Strategy by Ahrefs’ Mateusz Makosiewicz, HubSpot’s apical integrated competitors

Image Source

While we are fortunate to person a precocious domain authority and thrust millions of visits per month, having a blog of this size does not travel without challenges.

As our room ages, nan magnitude of opportunity for caller contented crossed our blog properties and clusters shrinks.

So, we decided to audit our library to find opportunities for optimization.

We hypothesized that we could uncover “greenspace” and “quasi-greenspace” — topics that we person covered but haven’t capitalized connected that good — by auditing nan oldest 4,000 URLs successful our library.

Although this was only astir a 3rd of our contented library, we believed we’d beryllium capable to unearth immoderate postulation opportunities and springiness our blog a boost.

Around nan aforesaid time, we started to consciousness the effects of Google’s March 2023 Core Update that emphasized experience, which our Technical SEO squad instantly started addressing.

However, different portion of that algorithm update emphasized contented freshness and helpfulness. In different words, really cutting-edge and useful our contented is to our readers.

This is wherever we really felt a consciousness of urgency.

Because we had 4,000 URLs pinch published dates ranging from 2006 to 2015, we already knew that this chunk of contented was not caller aliases helpful.

So, we sewage to activity and audited those blog posts complete nan people of 10 weeks.

Eventually, we added phases 2 and 3 to our scheme truthful we could further reside unhelpful contented and clusters.

How We Audited Our Oldest Content

how nan HubSpot Blog squad audits content; specify goals, stitchery contented inventory, retrieve data, measure content, and urge action

1. Define our goals.

Before we started auditing nan content, it was important for america to find nan objectives.

For immoderate publishers, nan extremity of a contented audit whitethorn see improving on-page SEO, enhancing personification engagement, aligning contented pinch trading goals, aliases identifying contented gaps.

For this peculiar audit, it meant uncovering “greenspace” and “quasi-greenspace” successful our blog library, and improving our wide contented freshness.

We besides had to find nan scope of our audit.

There’s nary correct aliases incorrect measurement to attack this. Depending connected your goals and nan size of your website, you could audit nan full point successful 1 go.

You could besides commencement pinch a mini information of your tract (such arsenic merchandise pages aliases circumstantial taxable clusters) and build retired from there.

Since HubSpot has specified a ample contented library, we opted to limit this audit to our oldest 4,000 URLs. Not only was this much manageable than reviewing each of our contented successful 1 audit, but this besides targeted URLs that were much apt to use from an update aliases prune.

We besides did this knowing that we would later reside nan remainder of our room during phases 2 and three.

2. Gather our contented inventory.

Once we established our goals and scope, we had to stitchery nan oldest 4,000 blog posts and put them into a spreadsheet.

This process tin alteration depending connected nan devices and CMS you use. Here’s really we did it utilizing Content Hub:

1. Log into HubSpot and navigate to nan Blog page successful Content Hub.

2. Navigate to nan Actions drop-down and click Export blog posts.

3. Select File format and click Export. This will nonstop each of your blog station accusation to your email. You will besides get a notification successful HubSpot erstwhile your export is ready.

4. Download your export and unfastened it successful your preferred spreadsheet package (I’m usually a Google Sheets girlie, but I had to usage Microsoft Excel since nan record was truthful large).

5. Review each file successful nan spreadsheet and delete nan ones that are not applicable to your audit. We instantly deleted nan following:

  • Post SEO title
  • Meta description
  • Last modified date
  • Post body
  • Featured image URL
  • Head HTML
  • Archived
6. Once nan irrelevant columns were removed, nan pursuing remained:
  • Blog name
  • Post title
  • Tags
  • Post language
  • Post URL
  • Author
  • Publish date
  • Status

7. Filter nan Post language file for EN posts only. Delete nan file erstwhile nan expanse is filtered.

8. Filter nan Status file for PUBLISHED only. Delete nan file erstwhile nan expanse is filtered.

9. Filter nan expanse by Publish date from oldest to newest.

10. Highlight and transcript nan first 4,000 rows and paste them into a abstracted spreadsheet.

11. Name nan caller spreadsheet Content Audit Master.

If you’re emotion fancy, you tin besides create a civilization study successful Content Hub and prime only nan fields you want included successful nan audit truthful you don’t person to select arsenic overmuch erstwhile mounting up your spreadsheet.

3. Retrieve nan data.

After compiling each of nan contented needed for our audit, we had to cod applicable information for each blog post.

For this audit, we kept it beautiful simple, only analyzing full integrated postulation from nan erstwhile almanac year, full backlinks, and full keywords.

We did this because our recommended actions for each URL were wished during post-evaluation. (We’ll screen this successful nan adjacent step.)

We obtained integrated postulation information from Google Search Console and utilized a VLOOKUP to lucifer each URL pinch its corresponding number of Clicks.

Then, we sewage backlink and keyword information by copying and pasting our audit URLs into Ahrefs’ Batch Analysis tool and exporting nan information into our spreadsheet.

At nan clip of our audit, nan Batch Analysis instrumentality could only analyse up to 200 URLs astatine once, truthful we had to repetition this measurement 20 times until we had information for each URL.

Luckily, Ahrefs has rolled retired a Batch Analysis 2.0 instrumentality since then, which tin analyse up to 1,000 URLs astatine once. So, if we were to do a akin audit successful nan future, it would return overmuch little clip to retrieve this data.

4. Evaluate nan content.

Next, we assessed each portion of contented by utilizing nan collected data. Then, we evaluated nan station itself to find nan following:

  • Type of Content
  • Freshness Level
  • Organic Potential

Type of Content

The HubSpot Blog is location to galore different types of blog posts, each serving a unsocial purpose. Labeling each station helped america find its relevance and became a cardinal facet successful our determination to update aliases prune.

While this isn’t an exhaustive database of each nan contented types you could find connected nan HubSpot Blog, we narrowed it down to nan pursuing for nan purposes of this audit:

  • Educational: A taxable that tin amended nan personification connected a symptom constituent aliases problem they cognize they have.
  • Thought Leadership: A taxable that tin amended nan personification connected a symptom constituent aliases problem they didn’t cognize they had until an master drew it to their attention.
  • Business Update: A HubSpot-related portion of news aliases a property merchandise that is apt not evergreen.
  • Newsjacking: An industry-related portion of news aliases a property merchandise that is apt not evergreen.
  • Research: A postulation of information aliases nan results of an research that is utilized to amended nan reader. This taxable whitethorn aliases whitethorn not beryllium evergreen, but nan contented is not and needs updates to enactment fresh.

Freshness Level

Because nan posts successful this audit hadn’t been updated successful a agelong time, nary of them could beryllium considered 100% “fresh.” However, we took different types of freshness into relationship erstwhile determining what action needed to beryllium taken connected nan URLs.

For example, immoderate topics, specified arsenic Google+, are so outdated that an update would beryllium silly. However, plentifulness of topics were still evergreen, moreover though our contented was not.

The pursuing standard helped america make decisions connected whether nan URL had worth pinch respect to freshness:

  • Outdated: The taxable is outdated, and an update whitethorn not beryllium possible.
  • Stale: The taxable is evergreen, but it would request an extended update to make nan contented much competitive.
  • Relatively Fresh: The taxable is evergreen, and it would only request a mean update to make nan contented competitive.

Organic Potential

To find nan integrated imaginable of each URL, we had to inquire ourselves nan pursuing question: Will anyone hunt for nan contented connected Google?

  • Yes: Someone would decidedly hunt for this, truthful we’ll request to optimize/recycle nan content.
  • No: Someone would not hunt for this. There’s nary constituent successful optimizing/recycling nan contented since there’s nary imaginable attraction keyword.

For each of nan posts marked “Yes” for integrated potential, we recommended a attraction keyword for nan re-optimized contented to compete for. We did this by evaluating nan existing title, slug, and content. Then, we did immoderate keyword investigation connected Ahrefs and reviewed nan Google SERP for that query.

We besides included nan attraction keyword’s monthly hunt measurement (MSV) to thief prioritize which updates to execute first. We did this by plugging nan recommended keyword into Ahrefs’ Keywords Explorer and adding nan MSV to our maestro sheet.

For an other furniture of caution, we besides checked for cannibalization connected each posts marked “Yes” for integrated potential. There are a fewer ways to do this:

  • Do a tract hunt and spot if immoderate URLs travel up for nan attraction keyword.
  • Plug nan attraction keyword into Google Search Console to spot if immoderate URLs travel up.
  • Plug nan attraction keyword into Ahrefs’ Keyword Explorer, scroll down to Position History, hunt your domain name, and select for Top 20 and your desired clip framework (I usually cheque nan past six months). If aggregate URLs are found, that whitethorn bespeak cannibalization.

If nan attraction keyword was flagged for cannibalization, we either recovered a different attraction keyword aliases noted that nan URL should beryllium redirected to nan fresher post.

If nary cannibalization was found, past we had nan greenish ray to move guardant pinch updating nan post.

5. Recommend an action.

Once a station was wholly evaluated, we turned nan insights into action items.

Each URL was placed into 1 of nan pursuing categories:

  • Keep: No action is needed because some nan contented and nan URL are good.
  • Optimize: The contented is bully but outdated successful position of freshness aliases SEO practices. Keep nan tone of nan article, but refresh and re-optimize to amended performance.
  • Recycle: The contented is not salvageable, but nan URL still has worth (in position of backlinks aliases keyword opportunity). Create caller contented from scratch, but clasp nan URL.
  • Prune: Neither nan contented nor nan URL has worth from an integrated standpoint.

Audit Insights

Out of nan 4,000 URLs we audited, 951 (23.78%) were categorized arsenic posts pinch integrated imaginable and recommended for optimization aliases recycling. Additionally, 2,888 URLs were recommended to beryllium pruned. That’s astir 72.2% of nan audit.

These posts either did not person integrated potential, posed a cannibalization risk, aliases were truthful outdated that location was nary constituent successful updating them.

The remaining 161 URLs either did not require immoderate action aliases had already been redirected.

How We Took Action

The action taken for a URL was wished by its imaginable for integrated traffic.

The URLs pinch integrated imaginable were delivered to our Blog squad and recommended to beryllium optimized aliases recycled.

Meanwhile, nan URLs pinch nary integrated imaginable were delivered to our SEO squad and recommended to beryllium archived aliases redirected.

First, let’s locomotion done really we took action connected nan posts recommended to beryllium optimized aliases recycled.

Taking Action connected Content pinch Organic Potential

Before addressing immoderate of nan 951 posts pinch integrated potential, we needed to fig retired nan following:

  • Our capacity for strategical study and little writing
  • The capacity of our in-house penning unit and disposable freelancers
  • Our capacity to edit nan updates

We coordinated pinch stakeholders and wished we only had nan bandwidth to update 240 posts successful 2024 (in summation to nan dozens of blog posts we update each month). This inaugural was internally known arsenic nan “De-Age nan Blog Project” and was led by my EN Blog Strategy teammate Kimberly Turner.

Once we knew really galore posts we could return on, we had to constrictive down which ones to prioritize. We did this by evaluating nan complexity of nan assistance required for each station update:

  • Simple Update: The contented updates needed are comparatively light, making them suitable for freelancers.
  • Complex Update: The contented updates needed are heavy, making them amended suited for in-house writers.
  • Recycle: Content is not salvageable, but nan URL is. Rewrite nan station from scratch, but clasp nan URL.
  • No Opportunity: Pass connected updating.

We primitively prioritized updating nan simplest URLs first, but later pivoted our strategy to tackle nan URLs pinch nan highest MSV potential, sloppy of update complexity.

We did this because we wanted to get nan astir we could retired of our updates.

De-Age nan Blog Results

Initially, we projected that these updates would beryllium complete by nan extremity of H1 2024, but we had to displacement our strategy … again.

Like galore different publishers, we felt nan effects of Google’s March 2024 Core Update arsenic good arsenic nan introduction of AI Overviews.

After having placed nan De-Age nan Blog Project connected clasp while we addressed nan issues, we deprioritized nan task wholly successful favour of higher-impact workstreams.

SEO, americium I right? It ever keeps you connected your toes.

Despite sunsetting nan task earlier it was complete, we were still capable to execute 76 station updates. Six months aft nan updates were implemented, nan cumulative monthly postulation for these posts had accrued by 458%.

This goes to show that moreover updating a mini information of URLs tin make a large difference.

Taking Action connected Content pinch No Organic Potential

While nan De-Age nan Blog Project was taking place, we besides took action connected nan 2,888 URLs that were recommended to beryllium pruned.

Since nan first audit didn’t see recommendations connected really to prune, we had to spell backmost and re-review each URL to find how we would prune.

Here’s really we evaluated nan posts:

  • Archive (404): The URL has little than 10 backlinks and nan backlink floor plan does not person value.
  • Redirect (301): The URL has much than 10 backlinks and/or nan backlink floor plan has value.

How precisely did we find nan backlink floor plan value? Rory Hope, HubSpot’s caput of SEO, recommended we travel these steps:

1. Login into Ahrefs and taxable nan URL into Ahrefs’ Site Explorer hunt bar.

2. Select Overview from nan left-hand sidebar.

3. Scroll down and click Backlink Profile.

4. Scroll down further and prime By DR under Referring Domains.

5. Analyze and analyse immoderate referring domains that are > 50.

6. Navigate to nan Referring Domain you’re investigating > 50 by clicking nan number.

7. Analyze nan Referring page.

Select Redirect (301) if:

  • The Referring page nexus is from a domain that still receives Domain traffic.
Select Archive (404) if:
  • The Referring page nexus appears to beryllium “spammy.” You tin find this by asking nan pursuing questions:
  • Does this website only people debased value impermanent station (SEO-led) contented from tons of different topics?
  • Does this website still people content? If not, disregard it.
  • The Referring page is from a website that you spot linking to a batch of EN Blog posts, done a RSS style automated linking system.

Additionally, each URLs branded “Redirect (301)” required a caller URL to beryllium redirected to.

When choosing a caller URL, we did our champion to prime nan astir applicable and akin page. If we couldn’t find one, we redirected to nan pillar page of nan cluster that nan station belonged to.

If for immoderate reason, nan URL didn’t beryllium to a cluster aliases location wasn’t a pillar page, we redirected it to nan HubSpot Blog homepage.

Decision-making for immoderate contented types was easier than others. For example, we were capable to automatically delegate 301 redirects to URLs that were flagged for cannibalization during nan first audit. We besides automatically assigned 404s to URLs pinch little than 10 backlinks branded arsenic Newsjacking and Business Updates.

Everything other was manually reviewed to guarantee accuracy. To make nan information process easier, we followed this determination tree:

the HubSpot Blog team’s contented auditing determination tree

It took my squad astir 2 and a half weeks to guarantee that each URL had nan correct label. In nan end, we had 1,675 URLs assigned to beryllium redirected and 1,210 URLs assigned to beryllium unpublished and archived.

Once each URL was evaluated, we were yet fresh to return action.

After coordinating pinch Rory and Principle Technical SEO Strategist, Sylvain Charbit, we decided to prune nan URLs successful batches alternatively of each astatine once. That way, we could amended show nan effect of redirecting and archiving a ample amount of content.

Originally, we planned to instrumentality our prune successful 5 batches complete 5 weeks, allowing america clip to show capacity during nan weeks successful between.

Batches 1 and 2 contained URLs meant to beryllium archived and unpublished, and batches 3 done 5 contained URLs designated for 301 redirects.

Because location were truthful galore URLs to unpublish and archive, we worked pinch developers connected HubSpot’s Digital Experience squad to create a book that would automatically unpublish and archive URLs and redirect them to our 404 page.

Then, we were capable to instrumentality nan 301 redirects pinch nan Bulk URL Redirect tool successful Content Hub.

Note: Although we were capable to activity done this process internally and decorativeness earlier our deadline, I want to admit that manually evaluating complete 2,000 URLs tin beryllium tedious and time-consuming.

Depending connected your resources and nan scope of your audit, you whitethorn want to see hiring a freelancer to thief your squad activity done a task this large.

Content Pruning Results

While we successfully implemented each batch, this process didn’t travel without a fewer roadworthy bumps.

Midway done our pruning schedule, Google rolled retired nan March 2024 Core Algorithm Update. We ended up placing our pruning schedule connected clasp truthful we could amended show capacity during nan update.

Once nan update was complete, we resumed nan remainder of our prune until it was complete.

Because of nan volatile hunt scenery successful 2024, we didn’t spot nan postulation gains we’d hoped to spot erstwhile nan prune was complete. However, we did observe a monolithic triumph for wide contented freshness connected nan blog.

At nan commencement of our audit successful 2023, we calculated nan freshness of our contented room by looking astatine each URL's people day and quantifying nan number of days since they were updated.

For example, opportunity nan existent day is November 12, 2024, and you person a station that was past updated connected February 19, 2008. Based connected nan 2024 date, nan station from 2008 is 16.7 years aged aliases 6,110 days.

Once we had each of nan ages for each station connected nan HubSpot Blog, we averaged those numbers to find nan mean property of our contented library, which was 2,088 days (5.7 years).

Since pruning 2,888 URLs (and updating hundreds of URLs from nan audit and beyond), nan HubSpot Blog's mean property has dropped to 1,747 days — that’s 341 days younger than erstwhile we started.

As contented freshness and helpfulness play an moreover greater domiciled successful hunt algorithms, being astir a twelvemonth younger tin make a large difference.

What’s Next?

Earlier successful this post, I mentioned that this audit is only 1 of 3 that my squad has worked connected successful 2024.

Our shape 2 audit focuses connected nan lowest-performing posts that were not included successful shape one, totaling complete 6000 URLs. Then, shape 3 assesses nan worth of our Blog’s taxable clusters.

We’re still taking action connected nan results from these audits, but I’m truthful excited to stock nan process and insights erstwhile they’re complete.

Ultimately, contented auditing is simply a occupation that is ne'er genuinely done — particularly erstwhile moving pinch ample libraries. You decorativeness 1 audit, past it’s connected to nan next.

Although nan activity tin beryllium tedious, nan rewards of improving contented quality, personification experience, and capacity make it worthy nan effort.