A twenty-three-year-old turns the Instagram feed into a fact stack—and calls the editorial voice barely opinionated.
Rohit Agarwala graduated with a BBA Honors degree, was no good at dancing or singing, and lived in 2020 lockdown watching peers form opinions on top of opinions on top of bad sources. Eighteen months in, Barely Opinionated is an Instagram-native news brand with two channels, a YouTube long-form arm, seven core staff and a fifteen-person extended bench, an eight-to-ten per cent engagement rate against the platform's four-to-six per cent baseline, and a refusal to take brand money that does not add value to the audience. This conversation is the operating manual for a Gen-Z newsroom that survives by curating other people's sourcing, not by chasing its own.
In sixty seconds.
Rohit Agarwala started Barely Opinionated in October 2020 because Gen-Z and young millennials were forming political opinions on top of other people's incorrect opinions on Instagram, and the Chinese-whisper output was scaring him. The product is curation, not reporting. A preliminary research team pulls the source material; Rohit and a head of research filter it for factual accuracy; the editorial wraps it in language a twenty-year-old will actually scroll past their gym reels for.
The voice is the brand. Barely opinionated means you read all sides, you mark who covered what, and then you tell the audience to make up their own mind. That is the explicit editorial discipline, and it is also why the brand has survived a politics beat that scares almost every other creator in the country. No clap-back. No notice. Factual sourcing on everything, no confrontation, every story shows which outlet covered which angle.
Eighteen months in, the operating shape is two Instagram channels (Barely Opinionated plus the spin-off Money Matters), a weekly long-form YouTube, four-to-five pieces a day, seven core staff and a fifteen-person extended bench, an eight-to-ten per cent engagement rate against the platform's four-to-six baseline, a research-cohort program that trains outsiders into "BO-certified" researchers, and a brand-deal filter that has cost real money but kept the trust intact.
Where to land in the conversation.
Each chapter opens the YouTube video at that timestamp in a new tab.
Five ideas to carry into your own work.
Mental models lifted from the conversation that travel beyond media. Each one is the kind of thing you can quote in an editorial meeting on Tuesday.
Barely opinionated.
The phrase is the editorial guardrail and the brand at once. The newsroom is barely opinionated, meaning it does not hold strong opinions, will listen to people, and will educate them about its own opinion if asked — but the default output is sourced facts with the outlet that covered each angle named on screen. The audience is then asked to make up its own mind. The output is curation, not commentary.
One-minute piece, hours of research.
The ratio Rohit names is brutal: hours of research underneath a one-minute or two-minute deliverable. A preliminary research team pulls sources; Rohit and a head of research go through them together; the piece is then structured in language an Instagram scroller will actually finish. The point of compression is not less work — it is more work invisibly absorbed.
The cohort as a recruitment funnel.
People kept asking to work for Barely Opinionated for free. Instead of saying yes one at a time, Rohit turned the demand into a three-month research cohort that trains accuracy, sourcing, and writing — and certifies graduates as "BO researchers" they can send into other brands. The cohort is a training product, a screening funnel, and a brand-extension move in one structure.
The brand-deal value filter.
Rohit has turned down "so much money" because the brand on offer added nothing to the audience. The discipline is explicit: only run integrations that the audience would call value-additive. The cost is real revenue declined; the asset is the trust premium that lets the next deal close at all.
Feed the algorithm, or stay invisible.
Rohit's read of the platform game is unsentimental. Creators who say "I don't care about the algorithm" are quietly losing. If you are on a social platform, you serve the algorithm and the audience together. The way out for an intellectual brand is to use the snackable layer to earn the addition-to-favourites that pulls more intellectual content into the user's feed. The hack runs through the algorithm, not around it.
Snackable to long-form, on the same brand.
One-minute Instagram is the door. Long-form YouTube is the room. Rohit calls long-form "the bread and butter" of trust-building, and the operating shape is one weekly YouTube on top of four-to-five short Instagram pieces a day. The economics of attention are the inverse of the economics of trust, and a serious creator runs both stacks at once.
Fifteen things to actually walk away with.
Each one carries the timestamps where the moment lives, and a transferable note for work that isn't media.
October 2020 was when the opinion stack collapsed.
Rohit dates the origin of Barely Opinionated to the lockdown months of October 2020. He was twenty-two, fresh out of his BBA Honors, with too much time on Instagram. What he saw on the feed was a structural collapse he names plainly — people in his generation were posting opinions based on incorrect facts; the next reader was forming their own opinion on top of that incorrect-fact opinion; the third reader was forming theirs on top of the second. Chinese whispers ending in gossip. Once Rohit said the word "gossip" out loud, the gap became a product.
The historical moment is worth marking. The same lockdown that made TikTok a cultural force in the West and Reels a cultural force in India also turned an entire generation into amateur political commentators with no editorial scaffolding. Most response strategies (fact-checking sites, traditional newsrooms launching apps) targeted the supply of misinformation. Rohit's instinct was to fix the demand — give the same scroller a product that satisfied the urge to feel informed, without the toxin.
BBA Honors, debating, and the school years he hated.
The biography Rohit gives is sparse and consistent. He went to National Public School in Bengaluru, did "hardly anything" through his school years, hated himself in school by his own account, and reset on day one of college by joining the debating society. The debating society is the closest thing he has to a journalism education — competition debates, geopolitics, economics, a habit of structuring a position from sources rather than feeling.
The point is not the pedigree, which is modest. It is that the editorial instinct of Barely Opinionated — show all sides, name who said what, refuse to commit to one side — is a debater's instinct, not a journalist's. The training that produced the voice is two years of college debating in Bengaluru, then a year and a half of "different places" of work, then the brand.
Seven core, fifteen extended, two heads of research.
Rohit's team shape, by his own description: seven core members and an extended bench of around fifteen people. The workflow has two named stations — a preliminary research team that pulls source material, and the editorial pair (Rohit plus another head of research) that filters for factual accuracy and shapes the structure. The rest of the headcount is design, video, camera, and editing. The operating ratio is roughly two researchers to one finisher.
The unspoken claim is structural. A traditional newsroom of twenty-two would have one editor and twenty reporters; Barely Opinionated has two filters at the top of a pipe that pulls from existing reporting elsewhere. The org chart is built for verification velocity, not for original sourcing. That is what lets four-to-five pieces a day ship with the engagement rate they hit.
Engagement at eight to ten per cent against a baseline of four to six.
The single number that does the most work in the conversation. Rohit says Instagram's average engagement rate, across creators, sits at four to six per cent. Barely Opinionated runs at eight to ten. The premium isn't from outrage; he names the comments section as the place he goes for delight, because the opinions there are "intellectually thought out" in a way he does not see elsewhere on the internet.
That makes the metric a derivative of the editorial filter. A page that picks fights gets reach. A page that picks facts and refuses to commit gets a comments section where the audience does the opining the brand refuses to do — and that argument-among-readers is what Instagram's engagement signal actually rewards. The product converts editorial restraint into a structural format advantage.
The Future Leaders audience: UPSC and CA aspirants.
Asked to describe the core audience, Rohit names a specific cohort: "Future Leaders of India" — curious, want to learn, preparing for UPSC, preparing for CA, the people he says will go on to be the actual builders of the country. The age band is 18-28; the entry point is Instagram; the use case is current affairs that they would otherwise need to assemble from PIB releases, The Hindu editorials, and a dozen YouTube channels.
The framing matters more than it looks. Most Gen-Z brands target consumption (fashion, gaming, lifestyle) or aspiration as image. Barely Opinionated targets aspiration as function — the people whose career path actually depends on knowing current affairs cold. That is a smaller market by ad-impression count and a much higher-trust market by lifetime engagement. The CPM curve runs the opposite way to the volume play.
The BO Research Cohort: turning inbound into infrastructure.
People kept reaching out asking to work for the brand for free. Rohit's response was to package the demand into a three-month training program — the Barely Opinionated Research Cohort — that teaches research, sourcing, and writing to outside applicants. Graduates become "BO-certified researchers" and are explicitly encouraged to go work for other brands and companies. The brand pays nothing in headcount and gains a distributed alumni network of trained-to-its-voice researchers across the media economy.
The structural cleverness is in the unbundling. A traditional newsroom would have absorbed the volunteers, paid them late, and lost the good ones in eighteen months. Barely Opinionated turned the same supply into a product (the cohort), a screening funnel (the certification), and a soft expansion of the brand into other newsrooms (the alumni). The same hours of staff time produce three outputs instead of one.
Money Matters: the same playbook with a different vocabulary.
Money Matters is Barely Opinionated's second Instagram channel, covering personal finance and money topics with the same editorial discipline. The sub-brand exists for two reasons Rohit is open about. The finfluencer space had a debate around SEBI licensing for finance commentators that Money Matters covered as a story; the audience around money content also pays attention to other money content, and the channel was a way to capture that flow without diluting the politics-and-current-affairs feed.
What is interesting is that Money Matters is sister-brand strategy, not category sprawl. The voice rules are identical — barely opinionated, sourced, refuses to commit — and the cost of running the second channel is the marginal cost of one more topic stream over a research stack that already exists. The same finishers, same designers, additional research-team slot. A creator economy almost always tries to spin off as a personality; the operating discipline at Barely Opinionated is to spin off as a category.
Why almost nobody else covers Indian politics on Instagram.
Vishal asks why other creators don't touch politics. Rohit's answer is direct: most are scared of the crackdown on media. Barely Opinionated has not had a clap-back, and Rohit attributes the immunity to two operating rules. First, every piece is non-confrontational — the editorial does not say this is good or this is bad, it says this happened and this happened, what's your opinion. Second, every claim is sourced and every story shows which outlet covered which angle on screen.
The implication for the wider creator economy is that the politics-shaped hole in Indian short-form media is not a content gap. It is an editorial-process gap. Creators who would happily cover politics if they had a verification pipeline that protected them have not built that pipeline; the ones with the pipeline (legacy newsrooms) have not built the Instagram-native format. Barely Opinionated sits in the intersection by deliberate operating design.
The farm-laws moment was the first real audience.
Asked for his favourite piece of work, Rohit goes back to the very early days of the brand, when the farm-laws protests in Haryana and Delhi were a national fracture and the misinformation around them was extreme. Barely Opinionated covered the episode by breaking down what each side was actually claiming, with sourcing. He says the brand picked up three or four thousand followers from the coverage, and what mattered more than the number was that both sides of the farm-laws debate were reading and thanking the brand — people who supported the laws and people who supported the farmers were both consuming the same pieces.
That is the test case for "barely opinionated" as an editorial line. If only one tribe is reading, the voice has collapsed into commentary. If both tribes are reading and saying thank you, the voice has held. The farm-laws coverage was the moment Rohit learned the voice was a product, not just a personal taste.
The Byju's piece and the Raghuram Rajan Dosa piece were the analytical breakouts.
Asked for the strongest performing pieces of content, Rohit names two that sit outside the political frame: one on Byju's (which he calls "the piece we wrote on Byju's") and one on Raghuram Rajan's Dosa Economics analogy. The pattern is consistent — anything on India's development, new policies, and structural economics moves the audience more than headline politics. The brand has the politics permission slip but the audience pays it in economics.
The lesson Rohit half-says: the readership Barely Opinionated has built is not actually a politics audience. It is a careful-explainer audience that uses the politics coverage as a quality signal, and then keeps reading on economic-development pieces that legacy newsrooms underprice. The brand's real category is structural-India explainer; the politics coverage is the credentialing.
Brand deals only if the audience would call them value-add.
Rohit has rejected what he calls "so many brands and so much money" because the integration would not have added value to the audience. The rule is explicit. Money on offer is not the qualifying signal; relevance to the reader is. He frames the surviving brand deals as "value-added brands to our audience, nothing else" — and points out, in the same breath, that the long-term ambition is to work with government agencies (income tax, traffic police, Bangalore civic information) precisely because that is what the audience would say is value-add.
This is a CPM strategy, not just a values statement. The trust capital that lets a brand deal close at all is the audience's belief that integration is not a betrayal. Every wrong brand deal is a permanent dent in the rate card; every right one compounds. Rohit is running the rate card defence by accepting smaller revenue today to protect larger revenue tomorrow — a discipline most creator-economy advisers will tell a 100k-followers founder to abandon.
The algorithm is the rent, and the rent is paid in format.
Rohit's view on the algorithm is unsentimental in a way that distinguishes him from a lot of older media operators. Creators who say they don't care about the algorithm are quietly losing if they don't have a separate distribution network. The algorithm collects data, decides what the user wants to see, and reaches the people who have already self-identified by behaviour. The way to use it on purpose is to make snackable content the door and intellectual content the room behind the door — the act of adding Barely Opinionated to favourites is what reshapes the user's feed toward more intellectual content over time.
The framing is platform-native. The algorithm is not a villain; it is the landlord. The rent is paid in format. The differentiation has to happen inside the format the platform monetises. The intellectual brand can win, but only if it speaks the algorithm's language at the entry layer.
The TikTok comparison that should worry a parent.
Vishal repeats an observation he has heard from American sources — TikTok in the United States serves a feed of partying and dancing; Douyin (the Chinese sibling app) serves a feed of AI demos and educational content to its under-18 users. Rohit confirms the framing as widely discussed and points to it as evidence of what he calls "psychological warfare" via algorithm — feeds tuned to make one population dumber, the other sharper, by deliberate curation choices upstream.
Whether or not one accepts the strong version of the geopolitical argument, the local implication is straightforward. The Indian Gen-Z feed is being shaped by foreign-owned recommender systems whose objective function is engagement, not informedness. Barely Opinionated is a small, deliberate counter-weight inside the same feed — a way for the user who has been served three hours of gym reels to also catch ninety seconds on the new direct-tax regime.
Azan, The Hub Bengaluru, and the co-founder who walked in through an EIR.
The co-founder story is small and well-told. Azan runs The Hub Bengaluru, a co-living plus co-working space in the basement of Safina Ventures. In 2021, deep in the lockdown, The Hub ran an Entrepreneur-In-Residence program — three months of free housing and workspace for early founders. Rohit applied, got in, and the two met in the program. Azan was already moving into media on the distribution side; he liked what Rohit was doing; the conversations turned into co-founder paperwork.
The structural detail behind the anecdote is that the co-founder did not come through the standard founder-matchmaking channels. He came through an in-person residency built by someone who was himself trying to figure out the Bengaluru media stack. The institutional layer (Safina Ventures, The Hub's EIR) acted as the dating site. That layer is invisible to founders who insist on "just network on Twitter."
Long-form is the bread and butter — trust is built where the scroll stops.
The Instagram piece is the door. The YouTube long-form is the room. Rohit is unambiguous about which is the "bread and butter": long-form. One-minute content gets the like and the swipe; long-form content is where the audience decides whether the brand is worth subscribing to in any deeper sense. The output rhythm is roughly one long-form YouTube per week against four-to-five short Instagram pieces per day.
The implication for the creator economy is that the right stack is both at once. Snackable for reach, long-form for trust. The brands that pick one are running a half stack. Barely Opinionated's hour-long podcast on whether 18-year-olds should pick science, commerce, MBA, BBA or certificate courses became one of the highest-trust pieces in the catalogue precisely because the audience could sit with the question for an hour.
The reach map already crosses Pakistan, Italy and the diaspora.
One of the smallest, most telling moments in the conversation. Rohit mentions doing an Instagram Live where Pakistani viewers showed up in the comments saying "love your content." Someone in Italy recommended Barely Opinionated to a friend. The brand was not built for a global audience and is not aimed at one, but the editorial discipline — sourced facts, no chest-thumping, no domestic-tribe signalling — travels naturally. The Indian-diaspora 18-to-28 audience is finding the brand without it being marketed at them.
The structural reading is that a neutral-frame Indian newsroom on Instagram has a second market hiding behind it: every person of Indian descent abroad who wants reliable current affairs without the toxicity. That market is invisible to legacy Indian newsrooms because they have not built for global feed surfaces. Barely Opinionated is the only thing on the shelf because nobody else is on the shelf.
Family, friends and the cost-of-entrepreneurship line.
Rohit's parents were skeptical when he first proposed Barely Opinionated. His father is a bureaucrat; his mother is a homemaker; his elder brother Mohit is at a startup called Smartin Spaces and is, in Rohit's reading, on the same entrepreneurial path. Rohit worked for a year and a half after college (he did sit for campus placement) before launching the brand, and the parental backing turned from hesitant to supportive once they saw the brand take shape.
The bit Rohit keeps coming back to is the family-time budget. He keeps two to two-and-a-half hours a day at the end of the workday for parents and friends, and explicitly reframes it as a debt owed to the parents who want time with their kids, not a indulgence the kid takes. "It's very selfish to think of it from our perspective" — the parents want their kids to spend time with them, and that flips the obligation. The line lands quietly in the conversation but it is the part of the worldview that explains the rest of it: a content brand built by someone who treats other people's perspective as the unit of work.
Lines worth keeping near your desk.
The jargon, unpacked.
Some of these will be obvious; some won't. Skim, mark the unfamiliar, come back later.
Check what you actually retained.
Try to answer before you click. The point is to notice where the conversation is fuzzy in your memory, then return to the transcript.
Five questions worth sitting with.
No correct answers. Type into the boxes — your responses are saved locally and exportable along with your notes.
Rohit chose neutrality as a wedge in a medium that rewards heat. Where in your own work would the opposite move — the disciplined refusal to commit — actually differentiate you?
The audience Rohit picked is people whose career outcome depends on knowing current affairs (UPSC, CA, 18-28 aspirants). Which audience for your product has career-outcome dependency on what you ship — and are you actually serving them?
Rohit turned down "so much money" to protect the rate card he could charge in year three. What revenue are you taking today that is silently capping your year-three pricing power?
Inbound exceeded capacity, so he built a cohort instead of hiring. Where in your operation does the inbound already exceed capacity, and could that demand become a school?
Rohit keeps two to two-and-a-half hours daily for family because the parents want time with the kids — not the other way around. What "selfish" framing of your own time would flip if you took the other person's perspective seriously?
Where to push back.
The strongest version of each disagreement, written to be persuasive — not to win.
"Neutral framing is its own bias — and a smaller one to call out."
The counter: there is no editorially neutral choice of which two angles to present, which outlets count as "sources," and which framings to leave out. Every act of curation is an act of opinion that is simply less visible than a hot take. "Both sides" framing has been thoroughly critiqued in mainstream media for normalising fringe positions and creating false equivalence — the same critique applies to a sourced Instagram carousel. The brand may feel neutral to the reader who shares its taste, but a Hindu nationalist reader and a left-secular reader probably both feel that Barely Opinionated leans the other way. That is the giveaway that the framing isn't neutral; it is centrist by default, and centrism is a position.
"Building a media business on Instagram is renting on quicksand."
The push: every major media brand that built primary distribution on a single platform has eventually paid for it — Facebook publishers in 2018, BuzzFeed across platforms, the entire mid-2010s pivot-to-video class. Meta has changed reach algorithms three times in the last five years; political content distribution has been quietly throttled in multiple jurisdictions; a single misclassified post can suspend an account holding a hundred-thousand-follower audience. The cohort and the long-form YouTube are insurance policies, but the centre of gravity is still on a surface Rohit does not own. The defensible answer is a newsletter list and a paid product the brand controls end-to-end — and the longer that move is deferred, the more the brand looks like a feature on Meta's platform rather than a business on its own.
"The trust premium does not survive contact with scale."
The counter: at a million followers, the editorial filter loosens by structural necessity. A 17-person production team cannot run hours-of-research on four-to-five daily pieces across two channels indefinitely; the cohort-trained researchers will absorb the editorial voice imperfectly; the comment section gets seeded by bots and partisan brigades whose presence corrupts the engagement-as-quality signal. The brand-deal discipline gets tested differently when an offer hits seven figures and the rejection is the price of a hire that would have made the next month's payroll. The trust premium is not a fixed property of the brand; it is a function of headcount, of process discipline, and of revenue desperation. Scale eats all three.
"AI-generated explainer content will commodify the BO format inside eighteen months."
The steelman: a competent foundation model with a retrieval-augmented pipeline can produce a sourced, neutral-framed, multi-outlet-cited explainer in two minutes at near-zero marginal cost. The hours-of-research moat that Rohit names is exactly the kind of labour that current generative systems are eating fastest. Inside eighteen months, the format will be a commodity; the only durable differentiation will be the editorial voice (which AI can mimic but not invent), the trust capital (which is brand, not format), and the audience relationships (which are owned data, not platform data). The brand survives only if it moves from "we make this faster than legacy media" to "we are the only one a thinking 22-year-old trusts," and that is a positioning move, not a production move.
Three angles on Monday morning.
If you don't run a Gen-Z newsroom, here's what to take.
If you're a founder
- Audit the pricing power you are giving up by saying yes to wrong-fit revenue today. The rate card in year three is set by the integrations you turn down in year one.
- If your inbound exceeds your capacity, build a school instead of hiring. Productise the demand. The cohort screens, certifies, and distributes the brand at the same time.
- Stack snackable and long-form on the same brand. The platform pays in reach; the audience pays in trust. Run both pipelines deliberately and price the output by which layer it earns you.
- Pick the audience whose career outcome depends on what you ship. Smaller market, higher attention, lower churn, larger lifetime value. The ad-impressions-per-user metric will look worse and the unit economics will look better.
- If you sell into a platform, accept the platform's grammar at the entry layer. The reformer who refuses to format for the recommender will be invisible. The differentiation has to happen inside the format the platform monetises.
If you're a content lead
- Make your editorial discipline a written rule, not a vibe. Barely Opinionated's rule is named on screen — "this happened, this happened, what's your opinion." Yours should be writable in one sentence.
- Distinguish your highest-engagement category from your highest-credential category. They are usually different products. Use the credential category to license the engagement category, and never confuse which is which.
- Build your verification pipeline before you build your audacious beat. The brand's safety on hot topics comes from process, not from posture.
- Run the test for whether ideologically opposed readers can consume the same piece and both feel served. If only one tribe shows up in the comments, the voice has slid into commentary.
- Track engagement-rate against the platform's category baseline, not as an absolute. The premium is a derivative of the editorial filter — and the moment the premium narrows is the moment to interrogate which filter has loosened.
If you're an operator or BD lead
- Run the brand-deal audit on relevance, not on cheque size. Every wrong integration is a permanent dent in the rate card; every right one compounds.
- Map the audience's career-outcome dependency on your output. If you can name what your reader needs to be on top of for their next promotion or exam, you can price into that need.
- Treat the recommender system on every platform you sell into as an editorial position, not a neutral utility. Design your counter-weight content as a deliberate add-back to the feed the platform's objective function produces.
- Get yourself into an in-person residency where adjacent founders are building adjacent stacks. The co-founder match Azan-Rohit happened through The Hub's EIR — not through cold DMs. Online networking under-delivers on this category of match.
- Plan the snackable-to-long-form journey for each piece. The Instagram reel is the door; the YouTube long-form is the room. Distribute the moments of trust deliberately across the two surfaces.
The brand's arc, briefly.
The story Rohit sketches, lined up against the landmarks of the Indian creator-media decade.
The whole conversation, searchable.
Click a timestamp to open YouTube at that moment. Click any line to highlight it (yellow). Highlights and notes save in this browser only.
00:00 in the page to seek.