SERPs.ioHelpful Content Update Recovery: What the Data Shows in 2026
On October 11, 2025, HouseFresh announced that its Google traffic had finally returned to pre-HCU levels. It had taken two years, one month, and a cross-industry policy change at Google to get there. The site had lost 95% of its search traffic in September 2023, and for 24 months it sat in the same queue as hundreds of other indie publishers, waiting for a core update that might reverse a classification nobody could officially confirm existed.
That recovery is a win. It is also an outlier. When Lily Ray examined 130 sites hit hardest by the September 2023 Helpful Content Update, 129 of them had only seen visibility decline since. Glenn Gabe tracked a larger cohort of around 400 sites that were "obliterated" by the HCU, and by August 2024 only 22% had recovered 20% or more of their lost traffic. The June 2025 core update improved those numbers. The December 2025 and March 2026 core updates created an entirely new cohort of affected sites. The data is messy, but the shape of it is now clear enough to describe honestly.
Lily Ray examined 130 sites hit hardest by the September 2023 HCU. 129 of them had only seen visibility decline since.
This article summarizes what recovery from the HCU and the 2024-2025 core updates actually looks like in practice. Who comes back. How long it takes. What the recovered sites did differently. And what the data says about sites still sitting in the queue.
TL;DR: what recovery actually looks like
- A small minority of HCU-affected sites have recovered. Glenn Gabe's tracking puts meaningful recovery around 22% as of late 2024, with higher numbers after the June 2025 core update. Full recovery to pre-HCU levels remains rare.
- Timelines are long. For non-YMYL sites, partial recovery typically takes 2-6 months of sustained work. For YMYL sites, 6-12 months is standard, and some categories take longer.
- Recovery usually requires an algorithm update cycle. John Mueller was explicit: "These are not 'recoveries' in the sense that someone fixes a technical issue and they're back on track."
- The December 2025 core update hit affiliate sites at a 71% rate, YMYL and health sites at 67%, and e-commerce at 52%, per analysis from multiple SEO firms covering the rollout.
- Sites that recovered share a pattern: genuine experience signals, brand building beyond search, reduced content volume with higher depth, and in many cases, nothing changed except Google's classification of them.
The honest version is that recovery is real, but uncommon, slow, and partially outside the site owner's control. The rest of this article breaks down why.
Glenn Gabe tracked around 400 HCU-obliterated sites. By August 2024, only 22% had recovered 20% or more of their lost traffic.
A brief primer: HCU, then core, then core again
The Helpful Content Update was announced in August 2022 as a site-level classifier that would demote content written primarily for search engines. The first HCU rollout in August 2022 was modest. The September 2023 HCU was not. Indie publishers, affiliate sites, and smaller editorial outlets reported 50-95% traffic losses, with sites like HouseFresh and Retro Dodo becoming public casualty examples.
In March 2024, Google absorbed the helpful content system into its core ranking algorithm. The separate HCU announcements stopped. The signal did not. Google described the March 2024 core update as one of the largest it had ever launched, with multiple reinforcing systems, and said the combined updates produced 45% less low-quality, unoriginal content in results.
After that, the cadence became roughly quarterly. August 2024 core update. November 2024 core update. March 2025. June 2025. December 2025. March 2026. Each update reshuffled winners and losers, and for sites sitting on a negative HCU-era classification, each one was a possible exit ramp. Most of them were not.
The December 2025 core update was the largest of the cluster. SEO firms reported that 15% of previously top-10 pages lost their rankings, with overnight traffic drops in the 40-70% range across affected sites. Affiliate sites were hit hardest, at a 71% rate. YMYL and health sites followed at 67%. E-commerce landed at 52%. The pattern echoed the original HCU: sites offering synthesized or thin content dropped, while sites with demonstrable first-hand experience held or gained.
The December 2025 core update hit affiliate sites at a 71% rate, YMYL and health at 67%, and e-commerce at 52%.
Casualty data: how the hit breaks down by category
Recovery rates look very different depending on what kind of site you run. The table below synthesizes data from Glenn Gabe's HCU tracking, Lily Ray's analysis, Semrush's December 2025 rollout study, and SEO-agency reports covering the 2024-2025 core update sequence.
| Site category | Share hit by Dec 2025 update | HCU recovery rate (2024-2025 data) | Typical recovery timeline |
|---|---|---|---|
| Affiliate and product-review | ~71% | 10-22% partial recovery, full recovery rare | 6-18 months, often requires structural change |
| YMYL (health, finance, legal) | ~67% | 15-20% partial recovery | 6-12 months, longest cycles |
| E-commerce | ~52% | 25-35% partial recovery | 3-9 months |
| Editorial and news | ~45% | 20-30% partial recovery | 4-12 months |
| SaaS and B2B blogs | ~35% | 30-40% partial recovery | 2-6 months |
| Independent publishers (indie) | ~60% | 15-25% partial recovery | 12-24 months, many never recover |
A few notes on reading these numbers. "Recovery" here mostly means partial recovery to 30-60% of pre-update traffic. Full recovery to pre-HCU baselines is still rare outside of specific cases like HouseFresh, which needed both internal work and external policy changes at Google to fully bounce back. "Hit rate" is the share of sites in a category that lost noticeable rankings during the December 2025 rollout, not the total share of that category on the web.
Affiliate sites take the hardest and longest path. Many relied on manufacturer-spec summaries or rewritten reviews, and Google's post-HCU systems weight demonstrated hands-on experience heavily. A site built around compiled specs does not have an easy fix. The structural change required is closer to rebuilding the publishing model than editing existing content.
YMYL sites take long because Google applies tighter quality thresholds to health, financial, and legal content. Recovery estimates from multiple 2025 analyses put YMYL recovery at 6-12 months for substantial improvement, versus 2-6 months for non-YMYL. Inaccurate health or financial content can cause real harm, and Google's algorithms treat the category accordingly.
B2B and SaaS blogs sit at the better end of the distribution. The content mix on these sites tends to be closer to documented practitioner experience (product tutorials, engineering write-ups, case studies) than to synthesized informational content. The structural signals Google now weights are often already present.
Recovery timeline: the numbers behind the wait
Google has been explicit that recovery is not a quick fix. John Mueller stated plainly that recovery is not "someone fixes a technical issue and they're back on track." Some reassessments take months. Some require a subsequent update cycle. Some never happen.
The table below consolidates timeline benchmarks from multiple 2025-2026 recovery guides and case studies.
| Recovery phase | Non-YMYL sites | YMYL sites |
|---|---|---|
| Weeks 1-4 | Minimal visible movement | Minimal visible movement |
| Weeks 5-8 | 5-15% partial recovery on best pages | Still minimal |
| Months 3-4 | 20-40% partial recovery | 10-20% partial recovery |
| Months 5-6 | 50-80% recovery possible | 20-40% partial recovery |
| Months 6-12 | Continued improvement, some sites exceed baseline | 50-80% recovery possible |
| Beyond 12 months | Plateau or slow compounding | Still on recovery path |
One frequently cited case from the 2024 core update cycle describes a site that recovered 55% of lost traffic by week 12 and exceeded its pre-update baseline by 15% at week 16. That curve is achievable but on the fast end. Most sites never reach the "exceeded pre-update" milestone. The pattern is more often a partial return that stabilizes below where the site was before.
There is a specific psychological dynamic that comes with this timeline. Publishers describe it as the "waiting-for-next-core-update" cycle. You do the work. You wait. The next update rolls out. Nothing changes. You wait for the next one. One analyst who wrote extensively about this pattern noted that after the August 2024 core update briefly lifted some HCU sites, later updates partially reverted those gains, compounding the frustration. The June 2025 core update finally produced more durable recoveries for a subset of the original HCU cohort, but only after nearly two years of stagnation for most sites.
For affected sites, monitoring daily ranking volatility across key terms is the clearest early signal of reclassification during a core update rollout; this is the exact use case rank-tracking tools like serps.io exist for.
Sites that recovered: what Search Engine Journal found
In February 2026, Search Engine Journal published Marie Haynes' analysis of four sites that recovered from the December 2025 core update. The sites were long-term clients, past clients, or sites Haynes had reviewed. She was careful to note that causation is difficult to prove with core updates, but the patterns across the four are worth reading together.
The common thread was trust. Not trust as an abstract E-E-A-T score, but trust as a set of observable brand-level signals: improved customer service perception, reduced complaint volume, better review sentiment, clearer authorship, and demonstrated hands-on involvement with the products being reviewed. One of the recovered sites was a medical e-commerce operation that had been hit by the August 2024 core update and worked for over a year on trust-layer changes before the December 2025 update released the suppression.
Haynes' broader finding was that the December 2025 core update shifted how Google evaluates demonstrated experience. It is no longer sufficient to say you have used a product. Google now weights signals that suggest the content creator has actually interacted with the subject matter. Specificity, details, original media, and cross-site consistency all feed into that evaluation. This is the same pattern we described in our analysis of experience signals and AI citations: the moat is proof of having done the thing, not polished prose describing it.
Other recovered-site patterns from 2024-2025 reporting include:
- Reducing ad density, particularly scroll-fixed ads and autoplay video. Zyppy's 50-site case study found this correlated with recovery more consistently than content changes.
- Brand-building activity outside search. Multiple recovery cases, including HouseFresh, involved YouTube channel growth and collaborations with trusted creators in the same space.
- Consolidating thin content. Pages that ranked before the HCU but offered 500-800 words of generic coverage often got merged into longer, more authoritative resources or removed entirely.
- Adding author infrastructure. Proper author pages, verifiable external profiles, and named byline consistency across the site.
Sites that have not recovered: the HouseFresh and indie-publisher pattern
HouseFresh is the best-documented indie casualty, and the most-cited recovery. Before the HCU, it was a small independent review site with about 4,000 daily Google visitors. After September 2023, that number dropped to 200. Gisele Navarro, the managing editor, wrote publicly about being outranked by affiliate-heavy content from major publishers using parasite SEO and content swarming tactics.
The HouseFresh recovery took two specific things working in tandem. First, Google enforced its site reputation abuse policy more aggressively from November 2024 onward, issuing penalties to major publishers including Forbes, Wall Street Journal, Time, and CNN. That reduced the competitive pressure from parasite SEO content. Second, HouseFresh invested in YouTube and a collaboration with Linus Tech Tips, building brand signals outside Google that eventually fed back into search. Navarro's own summary of the recovery was "hard work and a lot of luck."
The sites that have not recovered share a different pattern. Many of them look like HouseFresh did in 2024, minus the brand-building lever. They are small publishers with genuinely first-hand content, often written by the owner, covering a narrow topic area. Retro Dodo, niche travel publishers, single-operator review sites. Digitaloft's travel-publisher analysis found that 78% of 671 travel publishers lost organic traffic between August 2022 and March 2024, and most have not seen full recovery since.
The explanation Glenn Gabe has offered repeatedly is that HCU moved expertise signals from page-level to site-level evaluation. A single great page on a site classified as unhelpful still gets suppressed. A thousand great pages on that same site do not escape the classifier by volume alone. The lift requires Google's systems to re-evaluate the site as a whole, which typically only happens during a core update cycle and often requires observable changes that Google's algorithms can weight (brand mentions, author signals, original data) rather than just more content.
What the recoverers did differently
Looking across the recovered-site case studies from 2024-2026, the actions that correlate with recovery fall into four buckets.
Experience depth replaces breadth. Instead of publishing more content on more topics, recovered sites published less content that went deeper on topics the site already had standing in. This matches the December 2025 update's evaluation shift: Google is now weighting "content necessity" alongside quality. Sites that asked "Why should this content exist?" rather than "How can I make this better?" tended to recover faster.
Original media and data. Screenshots, photos, video, original testing data. The Wellows research cited in most recovery literature showed multimodal content with original images achieves a 156% higher citation selection rate in AI search, and the same signals that drive AI citations now drive core-update recovery. Stock imagery and generic illustrations do not move the needle.
Brand signals outside Google. The HouseFresh YouTube channel. Podcast appearances. Cross-posting on Reddit and LinkedIn. Named-author profiles on external industry publications. These are the signals Google uses to corroborate experience claims, and they take months to build. Most of the full-recovery cases in 2025 involved some version of brand-building that happened in parallel with on-site changes. The related pattern we wrote about in topical authority and AI rankings applies here too: cross-source consistency matters more than any single-page change.
Structural reduction, not addition. Multiple recovery playbooks emphasize what to remove. Ads fixed on scroll, autoplay video, thin content with duplicate intent, unnecessary affiliate links, AI-generated content published at scale. The Zyppy 50-site study explicitly flagged ad reduction as a recovery correlate, and multiple SEJ-reported cases involved content pruning. The sites that added more content without cleaning up the old content rarely recovered.
There is a fifth pattern that deserves its own mention: many recovered sites did nothing at all. Glenn Gabe's July 2025 observation was that most HCU sites showing improvement during the June 2025 core update had not made meaningful changes. They kept publishing at lower frequency, maintained the site, and waited. Google's systems re-evaluated and reclassified. This is the Mueller framing: recovery is not fixing something; it is Google changing its mind.
What recovery feels like from inside
The case studies describe mechanics. The Reddit threads, forum posts, and publisher essays describe the experience. Two years of sustained ambiguity. No notification from Google. No confirmation that a classifier is applied. No clear signal from the algorithm updates whether the work is helping. Just traffic graphs that plateau at 10-30% of their former levels and a quarterly wait for the next core update.
Publishers describe a common sequence. The update hits. You audit everything. You implement the recommended changes. You wait for the next core update. Nothing happens. You read the next round of analyst commentary, notice patterns you missed, implement more changes. You wait again. Months pass. Your revenue drops. You reduce freelance budgets. You stop hiring. You consider shutting down. Then one of two things happens. Either a core update produces a meaningful lift, which might or might not hold, or it does not, and you keep waiting.
This is the dynamic that makes HCU recovery different from recovering from a manual action or a technical issue. There is no ticket to open, no reconsideration request to file, no clear milestone to work toward. The signal is that work is necessary but not sufficient. It is required, and then the algorithm decides.
A practical framework for sites currently affected
The recovery literature converges on a reasonably consistent sequence. It is not a guarantee. It is the pattern that correlates with the recovered subset of sites.
1. Diagnose honestly. Confirm the drop aligns with an HCU or core-update date. Rule out technical issues first. Run the site through Google's own quality rater questions and answer as a stranger would, not as the owner. If you score yourself highly on every dimension, you are not diagnosing honestly.
2. Identify content necessity gaps. For each piece of content on the site, ask whether it adds something that does not already exist in higher-authority form elsewhere. Pages that summarize information available everywhere else are candidates for consolidation or removal. This is the "content necessity" evaluation Google appears to be running in 2025-2026.
3. Prune and consolidate. Remove or merge thin content. Reduce keyword-cannibalization clusters. Redirect pages that no longer need to exist. This is uncomfortable because it feels like shrinking the site, but the data suggests it helps.
4. Add experience depth. Pick the topics where you have genuine first-hand involvement and deepen the content. Original testing, screenshots, data, specific numbers, named tools, outcomes. See our content structure guide for AI citations for specific extraction-friendly formatting patterns. The same structure helps core-update recovery.
5. Build author infrastructure. Real author pages, verifiable external profiles, cross-site publication history. Author attribution correlates with both AI citation rates and core-update resilience. If you publish under "Admin" or "Staff Writer," that is a signal worth fixing.
6. Build off-site brand signals. YouTube, podcasts, Reddit, industry publications, speaking, collaborations. The signals that corroborate experience claims are the signals that support recovery. This is the longest-lead-time item on the list. Start immediately.
7. Reduce ad and monetization density. Particularly scroll-fixed ads, autoplay video, and aggressive affiliate link placement. The correlation with recovery is strong enough in the case-study data that it should be on the list even if causation is imperfect.
8. Publish less, publish better. If you are currently publishing three articles per week, try one per week with triple the depth. The algorithmic signals that hurt HCU sites were tied to volume-without-depth patterns. Slowing down is sometimes the fastest recovery move.
9. Wait for a core update cycle. Google has said recovery usually requires a core update. The calendar matters here. Implement the changes well before the next rollout so the updated signals are stable when Google re-evaluates.
10. Keep a recovery journal. Document what you changed and when. If the next core update lifts you, the journal helps confirm what worked. If it does not lift you, the journal prevents the trap of randomly reversing changes because you are impatient. This is how publishers have avoided making the recovery path worse during the waiting periods.
This framework is what the recovered sites did collectively, pieced together from the public case studies and analyst commentary. It is not a formula for guaranteed return to pre-HCU traffic. Most sites will not get there, even following every step. The framework is what maximizes the probability within a distribution where full recovery remains rare and partial recovery is the realistic ceiling.
The broader picture: a different kind of search
The HCU and the core updates that followed were not isolated algorithmic adjustments. They were Google's attempt to raise the quality floor in a search environment that was being filled with AI-generated content faster than it could be filtered. The timing was not coincidental. The updates landed alongside the generative-AI content explosion, and the resulting SERP changes are what we covered in the shift away from traditional results.
What recovery looks like now is shaped by that larger shift. A site that recovers its organic traffic in 2026 enters a SERP where organic clicks are worth less than they were in 2023. AI Overviews appear on roughly a quarter of US desktop searches. Reddit and YouTube absorb a growing share of informational queries. Traditional publisher traffic is structurally lower than it was before the updates, and even recovered sites are recovering to a ceiling that is lower than their former ceiling.
That is not a reason to stop working on recovery. It is a reason to frame the work differently. Core-update recovery, AI-search visibility, topical authority, and experience signal density are all expressions of the same underlying evaluation: does this site demonstrate real involvement with its topics, in a way that Google's and the AI systems' classifiers can detect? The sites that answered yes on that question are the ones that are either recovering or were never hit. The sites that could not answer yes are still waiting.
The honest version of HCU recovery in 2026 is this. It is possible. It is slow. It requires work that would make the site better even if recovery never came. And for a meaningful share of affected sites, the recovery will not come in the form they hoped for. That is not encouragement or discouragement. It is what the data shows.
Frequently asked questions
Can a site recover from the Helpful Content Update?
Yes, but recovery is uncommon. Glenn Gabe's tracking of around 400 HCU-affected sites showed that only 22% had recovered 20% or more of their lost traffic by August 2024. The June 2025 core update produced additional durable recoveries, and HouseFresh fully returned to pre-HCU traffic in October 2025. Full recovery to pre-HCU baselines remains rare.
How long does HCU recovery typically take?
For non-YMYL sites, partial recovery typically takes 2-6 months of sustained work paired with a core update cycle. For YMYL sites covering health, finance, or legal topics, the standard range is 6-12 months. Many indie publishers hit in September 2023 waited nearly two years before seeing meaningful improvement, and some have never recovered.
What percentage of HCU-hit sites have recovered?
Around 22% of tracked HCU-affected sites had recovered 20% or more of their lost traffic as of August 2024, per Glenn Gabe's dataset. Lily Ray's earlier analysis of 130 hardest-hit sites found that 129 had only seen visibility decline since. Recovery rates improved modestly after the June 2025 core update, but full recovery remains a small minority outcome.
Did the March 2024 core update fix HCU sites?
No. The March 2024 core update absorbed the helpful content system into Google's core ranking algorithm, but it did not reverse the existing classifications. Google described it as one of its largest updates and reported that the combined updates produced 45% less low-quality content in results. For most HCU-hit sites, the March 2024 update did not produce meaningful recovery.
What should you do if your site was hit by the HCU?
Diagnose honestly against Google's quality rater questions, prune thin content, deepen pages where you have genuine first-hand experience, add proper author infrastructure, build brand signals outside Google, and reduce ad density. Then wait for a core update cycle, because Google has stated that recovery usually requires one. Document every change so you can tell what worked after a rollout.
Is the Helpful Content Update still running?
The HCU as a separately-named update no longer exists. Google absorbed the helpful content system into its core ranking algorithm in March 2024, and subsequent updates have been labeled core updates rather than HCU. The underlying signal that demotes unhelpful content is still active and continues to reshape rankings with each quarterly core update.