What makes a good listicle?

How do you constitute a listicle that performs good successful classical and AI Search?

In last week’s Memo, I dug heavy into 5.3 cardinal hunt results to spot if Google demoted listicles. 

Quick recap:

• Listicles are still everywhere. At slightest 1 existent listicle appeared successful the apical 10 SERP results for 55.1% of the 60,000 queries analyzed, and successful the apical 3 for 32.3%.
• Query intent is the biggest lever. Google shows 5.3x much listicles erstwhile the query explicitly asks for a group of options.
• No cosmopolitan demotion. Google shuffled who ranks: only 52.5% of queries kept the aforesaid position-one URL betwixt January and August.
• Publishers hit brands. Publishers supplied 46.8% of the highest-ranking listicles vs. 19.2% for brands/vendors, and won 54% of head-to-head matchups connected the aforesaid SERP.
• Results alteration sharply by vertical. Top-10 listicle beingness ranged from 42.0% (entertainment/gaming) to 67.2% (beauty/fashion); brand-owned listicles appeared connected 46.2% of B2B package SERPs, reflecting really heavy B2B leans connected the maneuver for AI search.
• The existent threat is indirect: title for attention. 92% of listicle SERPs carried Reddit aliases YouTube, and AI Overviews covered 83.7% of queries connected mean (93.4% successful B2B).

In Part 2 below, I’m answering the question: What makes a listicle execute well?

What higher-ranking listicles person successful common

I analyzed complete 15 variables crossed 3,000 listicles, but astir of them didn’t past the power checks, which progressive a fewer analyses to find whether the findings held aft normalizing definite variables. Interpret the ones pinch a “yes” successful the Direction stable file beneath arsenic “maybe, but we cannot beryllium it yet.”

The clearest awesome was freshness. A day from the erstwhile 2 years appeared connected 66.6% of pages successful the apical 3, compared pinch 57.3% of pages successful positions 8-10. That’s apt not a surprise: Search and LLMs emotion freshness.

An aged date, aliases a day the crawler could not interpret, appeared connected 8.2% of top-3 pages and 15.1% of pages successful positions 8-10. 

After controlling for vertical, query wording, and listicle type, an aged aliases unreadable day was associated pinch 56% little likelihood of ranking successful the apical 3 (when filtering down to 1 page per domain).

Counted listicles execute better. A counted listicle is 1 that has an definitive point count aliases numbered-list grounds successful its title aliases hunt snippet, specified arsenic “17 Best Project Management Tools.” An editorial roundup recommends aliases compares respective options without an definitive number, specified arsenic “The Best Project Management Tools.” The sanction describes really the page was classified successful this analysis, not whether personification manually counted each point connected the page.

• 56.8% of the nonstop matchups went to counted listicles. This comparison covers 6,114 hunt results pages wherever a counted listicle and an editorial roundup some appeared successful the integrated apical 10.
• The champion counted listicle averaged position 3.77 connected the 28,228 hunt results pages wherever 1 appeared successful the apical 10. The champion editorial roundup averaged position 4.59 crossed 10,791 hunt results pages. “Best integrated rank” intends the highest mean unpaid position reached by that page type connected each hunt results page.
• Counted listicles landed successful the apical 3 55.9% of the clip erstwhile coming successful the apical 10. Editorial roundups did truthful 41.4% of the time.
• The shape was not universal. Editorial roundups won the nonstop matchup successful user technology, fittingness and sports, and automotive, while counted listicles won successful the different 12 verticals.

Larger title numbers rank higher. The rank shape strengthens arsenic the starring number successful the title rises. Among definitive database queries, counted listicles pinch 51 to 100 successful the title had a mean champion integrated rank of 4.11. Titles pinch 2 to 5 averaged position 5.85.

The starring title number is only a proxy for database length. A title tin opportunity “50 ideas” while the page contains fewer, more, aliases otherwise grouped items.

Top-3 pages had 21.5% much words and 38.9% much images astatine the median, but these differences didn’t stay accordant successful each adjusted model. Median numbered-heading and ordered-list counts were zero successful some groups.

What doesn’t help: More conception headings did not foretell a amended position. A one-standard-deviation summation successful heading count aft adjusting for unusually ample values was associated pinch 29% little top-3 odds.

So, while signaling extent done large numbers successful the title earns higher rankings successful this dataset, over-structuring contented pinch galore headings is counterproductive. 

“50 champion tools” whitethorn awesome comprehensiveness, lucifer definitive database intent, aliases pull clicks. But 50 products tin besides look successful cards without creating 50 headings.

Self-serving listicles are good and alive

In Part 1, we saw that immoderate listicles mislaid traffic. In the cohort I analyzed for Part 2, I recovered that self-serving listicles that were published toward the extremity of past twelvemonth grew successful traffic: 3.5x much than the rest, successful fact.

So, while older cohorts were partially demoted by Google, newer ones recreation well. However (!), vendors putting themselves first successful their listicles is not responsible for their growth. They turn despite it. 

After accounting for starting traffic, industry, listicle type, existent rank group, and repeated pages from the aforesaid publisher, the advantage quickly disappeared. Self-promotion was associated pinch only 4% stronger growth, pinch a 95% interval ranging from 51% little to 120% higher (p-value was 0.91).

My publication is that companies put much effort into listicles aft realizing their domiciled successful AI Search and that’s why we spot the growth. At the aforesaid time, the consequence supports the statement that self-serving listicles still work, moreover though Google demoted a batch of them.

Details:

• Across 83 manually reviewed pages from 8 industries, median integrated postulation estimates reached 1,977 successful August 2026 compared to 568 for 600 power listicles.
• 74.7% of eligible B2B vendor comparisons put the publisher’s ain merchandise first. 62 classed the vendor first vs. 21 that did not.
• Of the 434 vendor-owned B2B listicles I reviewed, 83 compared the vendor pinch genuine adjacent alternatives.
• 90.4% of self-promoting pages gained measured postulation betwixt August 2025 and August 2026, compared pinch 67.3% of the wide benchmark.
• Among pages that already had measurable postulation astatine the starting point, the median summation was 186%, compared pinch 20% for different listicles.

Listing your marque first: Cringe?

I get the cringe facet of penning a listicle astir your marque and naming yourself first. Besides the mobility of whether humans would really publication this (or spell to Reddit alternatively because they’re alert of the bias), really do you morally reconcile this? 

My proposal to clients is to usage third-party information arsenic a impervious constituent for your listicle ranking. If you person G2 reviews that support you being “the champion for x”, you tin constitute astir it pinch a clear conscience. If not (and if you attraction astir this), possibly listicles are not the champion option. Maybe leaning connected an connection partner aliases creator to do this is the amended way.

There is an inherent philosophical problem pinch Google Search. When you hunt for a query for illustration [project guidance software] and you’re Asana, shouldn’t you automatically show up successful the hunt results?

The reply is no because location are only truthful galore package solutions and Google doesn’t want to beryllium “on the hook” for which brands show up first and why. 

The workaround: Instead, it often links to 3rd parties that curate brands for illustration publishers aliases marketplaces. Hence, listicles. Listicles are fundamentally curations and they debar Google having to return a stance.

Now, AI Overviews alteration that. In AI Overviews, you get a clear proposal for brands (see screenshot above). It’s a philosophical displacement successful really Google takes a position connected which brands to recommend. And portion of what allows this is personalization and longer (think: unique) prompts.

As a result, while listicles still rank good and mightiness pass AI answers (perhaps less, these days), Google’s displacement to sanction brands successful AI Overviews makes their worth moreover much questionable.

Methodology

This follow-up combines 4 analyses from the original study of 59,984 valid Google hunt results pages crossed 60,000 US English queries and 15 verticals.

First, I compared counted listicles pinch editorial roundups.

• A counted listicle showed an definitive point count aliases numbered-list grounds successful its title aliases hunt snippet.
• An editorial roundup recommended respective options without an definitive number.
• Direct matchups only included hunt results wherever some formats appeared successful the integrated apical 10.

Second, I crawled a balanced sample of 3,000 presently ranking listicles.

• The sample contains 200 pages from each vertical, divided evenly betwixt counted listicles and editorial roundups.

The crawl recorded nonsubjective page signals, including connection count, headings, estimated database items, images, writer information, dates, canonical URLs, and system data. It did not judge first-hand experience, merchandise testing, subjective contented quality, aliases monetization.

To place which signals were associated pinch higher rankings, I compared pages whose champion observed position was successful the apical 3 pinch pages successful positions 8 to 10. The adjusted models accounted for vertical, query wording, listicle type, really often a page appeared successful the study, and repeated pages from the aforesaid domain. I besides repeated the study pinch 1 page per domain to limit the power of ample publishers.

Third, I compared integrated postulation estimates betwixt January 15 and a unrecorded August 24, 2026 snapshot. The sample originates pinch pages that classed successful August and looks backward, truthful it describes today’s surviving pages alternatively than each listicles published during the period.

Finally, I reviewed vendor-owned listicles for self-promotion. A page qualified only erstwhile it compared the vendor’s merchandise pinch genuine adjacent alternatives. I classified it arsenic self-promoting erstwhile the vendor placed its ain merchandise first, branded it “Best Overall” aliases “Top Pick,” aliases different made it the first substantive recommendation.

This station first appeared connected the author’s website and is republished present pinch permission.