Subscribe
Share
Share
Embed
This week tracks why opens, likes, ad exposure, and video delivery numbers can mislead without platform context and infrastructure.
Media impact looks different when Kompas email clicks, short-form video bandwidth, gambling ad exposure, and platform-shaped reviews are measured beyond surface counts.
Covers 2026-10-14 to 2026-10-21; 5 free papers from 40 selected papers.
This Week in Media Measurement tracks research on how media, platforms, and marketing are measured, from social media and web analytics to campaign evaluation, audience behavior, AI-driven content, and privacy-preserving methods.
Episode covers 2026-10-14 – 2026-10-21.
Themes: social media, digital marketing, consumer behavior, mental health, media literacy, adolescents, purchase intention, digital media
Methods: survey, qualitative, quantitative, case-study, cross-sectional, content analysis
Premium also covers 9 related news stories, including martechcube.com — Generative Engine Optimization Is Now a Mainstream ..., everything-pr.com — 25 Perplexity Prompts for Media Research, and jakartaglobe.id — Indonesia Eyes Copyright Reform as Media Firms Struggle ...
The premium version of this podcast covers all 40 research articles and 9 news stories selected for the episode. Subscribe to the premium podcast.
Generated by paperboy.fm.
This Week in Media Measurement tracks research on how media, platforms, and marketing are measured, from social media and web analytics to campaign evaluation, audience behavior, AI-driven content, and privacy-preserving methods.
Subscribe for the premium version of this podcast: https://paperboy.fm/podcasts/media-measurement/subscribe
Jenny: Have you ever wondered how much of the video you scroll past was downloaded for no reason?
Davis: Constantly, because my phone acts like every half-second thumb twitch is a binding contract with a data center.
Jenny: Right, and I'm suspicious of any fix that says cheaper delivery with no catch, because video quality is where the catch usually hides.
Davis: But that's why measurement matters: in one short-form video experiment, capping bandwidth cut the median amount downloaded by 15% to 45%, while the main quality signals barely moved in most cases, so today we're asking what else our media metrics are quietly missing...welcome to This Week In Media Measurement on paperboy.fm.
Jenny: This week, we found 123 qualified papers out of 3,039 hits, with 339 unique authors across 46 countries. That's the scale: a little over one hundred usable studies, about three hundred forty authors, and a much wider map than last week.
Davis: The headline count barely moved: 121 to 123, up 1.7%. So the feed isn't exploding; surveys at 42 papers and qualitative studies at 31 are still doing a lot of the work, with quantitative studies at 26.
Jenny: But the search surface got noisier. Hits rose from 2,830 to 3,039, up 7.4%, while authors fell from 384 to 339, down about 12%, so what's driving more material with fewer named people?
Davis: Part of it looks geographic. Country coverage jumped from 27 to 46, a 70% increase, and the top countries are Indonesia with 10 papers, India with 9, and China with 8, so the week got broader without one giant author pool taking over.
Jenny: The author mix is almost thirds: 100 first-time authors, meaning first-ever papers in the metadata, 116 emerging authors, and 123 experienced authors. I'd be careful reading prestige into that, because we don't have institution data this week, and cities are blank.
Davis: The themes keep the episode's thread tight: social media leads with 27 papers, then digital marketing at 10 and consumer behavior at 8, with mental health at 6. That's the reminder that media measurement isn't just counting exposure anymore; platform design, AI systems, infrastructure, and context keep changing what those counts mean.
Jenny: Alright, let's get into the papers, and I'm starting with a very practical one: Nikolas Wehner and colleagues have a twenty twenty-six TOMCCAP paper called Demonstrating a Framework for KPI/KQI Analysis of Network Management on a Short-Form Video Use Case. It's about TikTok-style video, where apps pre-load clips so scrolling feels instant, but a lot of that downloaded video may never get watched.
Jenny: The surprise is that a network can cap bandwidth, meaning it limits how fast those video flows download, and still keep the viewer experience mostly intact. In their experiments, capping cut median download volume by fifteen to forty-five percent, and cut bandwidth excess by eighteen to fifty-two percent on average.
Davis: How did they decide the viewer experience was still okay, rather than just proving the network got cheaper to run?
Jenny: They measured both sides together: KPIs, or key performance indicators, are the network numbers like throughput and downloaded volume, while KQIs, or key quality indicators, are the user-facing signs of quality, like whether playback starts smoothly. They built a framework for that pairing, then tested bandwidth caps on three popular short-video services across different scenarios, and the quality indicators were nearly unchanged in most cases, with only minor statistical differences in a few.
Jenny: The caveat is real, though: this is experimental work on specific services, so I wouldn't turn it into a universal rule for every short-video app or every network condition. But as evidence, it's stronger than a dashboard hunch, because they're not just counting saved bytes; they're checking whether the app still feels usable.
Davis: That's a clean example of the infrastructure side of measurement. If platforms only look at engagement, they might miss waste; if networks only look at traffic, they might damage quality, and this paper says the useful answer is to measure both before you touch the throttle.
Davis: That last paper was about not trusting one dashboard number, like saved bytes, until you check the human side too. This one flips that into a much darker setting: Chris Vargo, Toby Hopp, and Harsha Gangadharbatla have a twenty twenty-six Policy & Internet paper called Gambling Advertising and Problem Gambling in the United States.
Davis: The plain finding is that more gambling ad spending in a state lined up with more searching for help with gambling problems, even when sports betting's legal status did not explain it. In Study one, they linked gambling ad outlays from twenty fourteen through twenty twenty-three to Google Trends data for one hundred sixteen help-seeking search queries, across forty-one states and one thousand six hundred forty state-quarters, meaning one state observed during one three-month period.
Jenny: But are we seeing advertising causing harm, or advertising showing up in places where gambling problems were already rising?
Davis: That's the central caveat, and the authors say it directly: these observational designs do not identify causal effects. What makes the evidence harder to wave away is the second layer, where they surveyed four hundred nineteen mobile-platform users and paired that with passive digital traces, meaning observed phone or web activity rather than just memory, and exposure to gambling ads on Facebook, TikTok, and X was associated with gambler status, gambling frequency, and whether someone made any observed gambling-site visit.
Davis: They also looked at clean gambling ads, basically ad exposures they could filter and classify with more confidence, and those were tied to higher twenty-four-hour odds of later gambling touchpoints, especially web visits. Among gamblers, the ad-triggered touchpoints tracked problem-gambling scores more consistently than raw ad volume, which is a very measurement-week result: the event after exposure mattered more than the count of ads by itself.
Jenny: So I wouldn't call this proof that one TikTok ad creates a gambling problem, but I would call it a safety signal that platforms and advertisers should not treat like a normal conversion funnel. It fits the platforms-shape-signals thread, because the same gambling message can look like marketing success in one dashboard and like help-seeking pressure in another.
Jenny: That point about the event after exposure mattering more than raw ad volume is a neat bridge, because this next paper has the same measurement instinct in a much lower-stakes setting. Anggun Aulia Putri and A. Fadillah look at The Effect of Visual Information Chunking-Based Marketing on Customer Engagement With Kompas One, which is a premium content bundle anchored on Kompas.id.
Jenny: The plain version is simple: people were already opening the emails, but the old text-heavy design wasn't getting them to click. When Kompas One switched to Visual Information Chunking, meaning breaking information into scannable visual blocks instead of dense paragraphs, click-through rate rose seventy-one point four three percent, with p equals zero point zero zero seven.
Davis: If opens stayed flat, what exactly changed after people got into the email?
Jenny: They compared thirty baseline text-heavy campaigns from September and October twenty twenty-five with thirty Visual Information Chunking campaigns from November and December, across randomized, non-overlapping cohorts of prospective, at-risk, and churned subscribers. Click-to-open rate, which asks among people who opened, how many clicked, rose sixty-one point five four percent, with p equals zero point zero zero eight, while open rate showed no significant change at p equals zero point eight zero two. The caveat is real though: this is one premium bundle and one email program, so I'd want replication before treating it like a universal design law.
Davis: That's very Beyond Vanity Metrics. The subject line got them in the door at about the same rate, but the room was easier to navigate once they arrived, so for subscriber marketing the useful dashboard isn't just opens; it's what the clearer layout lets people do next.
Paperboy.fm: This is the free version of the podcast. Subscribe at paperboy.fm to access a dozen different paper review podcasts for five dollars a month.