Arattai
Root cause analysis of Zoho's Indian messaging app that hit #1 in downloads then lost 99% of its user base within 8 weeks — a masterclass in what network-effect products get wrong.
The Problem
Despite hitting #1 in downloads during the WhatsApp privacy backlash, Zoho's Arattai lost 99% of its user base within 8 weeks due to a failure to sustain network effects.
Business Outcome
Diagnosed the failure of a viral app, offering a framework for sustaining network effects post-acquisition.
How I'm approaching this
I'm treating this as an RCA interview question: a product went from #1 to irrelevant in 8 weeks. My job is to figure out why — not just list all the bad things that happened, but identify the primary cause, separate it from the compounding factors, and explain what I'd have done differently. I'll show every step of my reasoning, including the hypotheses I considered and rejected.
Step 1 of 8
Clarify what I'm diagnosing
My first instinct when I hear "downloads dropped 99%" is: what exactly does that mean? Download count and active user count are very different metrics, and they have different root causes.
Downloads dropped — but what did the activation curve look like?
This is the question I'd ask before anything else. Downloads dropping 99% could mean: (a) people stopped installing, or (b) people installed but never used it. If it's (b), the root cause is in the product experience, not the marketing. My hypothesis going in: it's both, but the activation failure is more diagnostic.
Is this a consumer metric or a business metric problem?
Consumer. Arattai is a B2C messaging product and the failure is at the user level — people installing and not returning. This isn't a monetisation or revenue RCA. I'm diagnosing: why did users not stay?
What's the baseline I'm comparing against?
The peak download rate during the WhatsApp controversy window in January 2021. That peak was itself extraordinary — I'd want to know if the question is "why did it collapse" or "why didn't the peak sustain". My read: both. The peak was artificial, and the collapse was the inevitable consequence of an artificial peak with no retention foundation under it.
0st
App Store rank (peak)
India App Store, Jan 2021
0%
Download collapse
Within 8 weeks of peak
0M
WhatsApp India users
The switching barrier
0%
Est. D30 retention
No network = no return
Step 2 of 8
Understand the context before forming hypotheses
I'd spend time understanding the product and market before I hypothesize. An RCA on a product I don't understand is just guessing.
Arattai
The product
Consumer messaging app by Zoho Corp. Tamil: 'to chat'. Positioned as privacy-first WhatsApp alternative for India.
India B2C messaging
The market
WhatsApp is the de facto communication standard — 500M+ users, embedded in daily life for everything from family groups to business transactions.
Jan 2021
The moment
WhatsApp announces mandatory Facebook data sharing. IT Ministry summons WhatsApp. Vocal for Local sentiment is at peak post-China app bans.
🇮🇳 The Vocal for Local Context — Why I Can't Ignore It
India had spent six months in 2020 banning Chinese apps under Atmanirbhar Bharat. When the WhatsApp controversy arrived in January 2021, the IT Ministry publicly summoned WhatsApp and government officials encouraged the switch to Indian alternatives. Arattai was the right nationality at the right moment. I'd flag this early in my RCA because it means the download peak was inflated by two independent forces — the WhatsApp controversy AND a government-amplified nationalist sentiment — neither of which was product-driven demand. Any product with an Indian flag and a privacy story would have seen similar numbers. That's important context for what the peak actually means.
How the 8 weeks unfolded
India bans 59 Chinese apps including TikTok and WeChat. PM Modi's Atmanirbhar Bharat campaign gains momentum. Indian consumers are primed to favour homegrown digital products. This is the cultural precondition for Arattai's surge.
WhatsApp announces mandatory Facebook data sharing by Feb 8. The announcement triggers immediate backlash across Indian media, tech communities, and government circles. The event that creates the surge.
India's Ministry of Electronics and IT formally summons WhatsApp. Government officials publicly criticise WhatsApp and encourage citizens to switch to Indian alternatives — providing Arattai direct political tailwind.
National tech media profiles Arattai as the Vocal for Local answer. Government-adjacent Twitter handles share the app. The triple catalyst (controversy + nationalism + media) creates extraordinary acquisition.
Estimated 500K+ installs in the first week. The surge is real, but it's driven by sentiment, not product need. This is the number I'd want to interrogate: how many of those installs activated?
Elon Musk tweets 'Use Signal'. Signal's daily downloads hit 7.5M globally. The privacy-exodus narrative consolidates around Signal. Arattai, without a comparable endorser or trust story, loses the positioning battle.
WhatsApp delays enforcement to May 2021. The urgency evaporates. Govt officials stop amplifying. The civic reason to have Arattai installed disappears. Downloads collapse within days.
Arattai effectively disappears from public consciousness. The product that was #1 in the App Store weeks earlier has near-zero daily download rate. Nothing changed in the product. The narrative changed.
Step 3 of 8
Rule out the boring explanations
Before I hypothesize interesting causes, I'd rule out the obvious ones.
Conclusion
The collapse is real and external event-driven. The cause is not a product failure in the traditional sense — Arattai didn't break. It was never used. I'm now diagnosing: why did downloads collapse when the external catalyst disappeared, and why was there no organic floor?
Step 4 of 8
Map all the hypotheses
Before I pick a primary cause, I'd generate all plausible hypotheses. Click each bone to see my assessment — whether I think it's primary, secondary, or a compounding factor.
Click any hypothesis category to see my read on it
Step 5 of 8
How I'd rank and test each hypothesis
I have 7 hypotheses. I'd rank them by explanatory power — which one, if true, best explains the full shape of the collapse including the timing.
1Network effect cold-start — no one to message
Look at Day 1 and Day 7 retention rates. Specifically: what % of users who installed ever sent a message? If that number is <10%, the product never activated. Cross-reference: did Arattai have a 'contacts on Arattai' feature at launch? If not, the cold-start was designed in.
Verdict: Primary — explains the floor being zero
2Event-driven peak with no organic baseline
Look at the download curve before Jan 2021. Was there any organic growth before the controversy? If organic growth was near zero, the peak was 100% borrowed. The collapse is inevitable — you can't sustain borrowed demand.
Verdict: Primary — explains why the collapse was total, not partial
3Govt / Vocal for Local created civic installs not product installs
Compare download geography. If downloads spiked uniformly across India (not tech-hub concentrated), it suggests the driver was broad media/political amplification, not organic tech adoption. Civic installers have near-zero product intent.
Verdict: Strong contributing — explains the peak's magnitude
4Signal won the privacy narrative
Look at cross-install rates — did users who installed Arattai also install Signal? If yes, Arattai was a backup choice, not a primary one. Also look at: did Arattai's rank drop immediately after the Elon Musk 'Use Signal' tweet?
Verdict: Secondary — explains why stickiness was near zero even for motivated switchers
5Product feature gap (no calls etc.)
If the product gap were primary, I'd expect activation rates to drop after the first session as users discovered what was missing. But the retention problem likely appears before first session completion — users open it, find no one there, close it. Feature gap is downstream of the network problem.
Verdict: Contributing but not primary
Step 6 of 8
Trace each observation to its root
For each significant observable symptom, I'd walk the causal chain backward until I hit something structural. Click each to see the chain.
Step 7 of 8
What I'd stake on as the root cause
Primary hypothesis
Arattai launched a network-effect product into a cold-start problem with no cold-start strategy. The download peak was borrowed from an external event. When the event ended, there was no organic floor because the product had never activated.
The Vocal for Local sentiment and the WhatsApp controversy explain the extraordinary peak. The network effect failure explains why the floor was zero. These two hypotheses together account for the entire shape of the curve — the spike and the cliff — without needing any other cause.
Evidence that supports it
- →Download collapse correlates exactly with WhatsApp policy delay, not with any Arattai product change
- →No evidence Arattai had a contact import or 'find friends' feature at launch
- →Signal maintained meaningfully higher users post-peak — it had a pre-existing tech community to seed from
- →No organic baseline before the Jan 2021 surge — 100% borrowed demand
Strongest counter-argument
- →Even with a cold-start strategy, WhatsApp's inertia (500M users) may have been insurmountable
- →Signal also had a cold-start problem in India — why did it fare better?
- →Product quality (no calls etc.) may have been the real activation killer, not network absence
How I'd address the counter-arguments
WhatsApp's inertia is real but not deterministic — Signal and Telegram both built viable user bases in India. The difference is both had seeding strategies: Signal seeded through tech/privacy communities, Telegram through large-group enthusiasts. Arattai had neither. On Signal's cold-start: Signal's endorsement by Musk created a concentrated seeding event in a tech-savvy community that could afford to install en masse simultaneously. That's the rarest cold-start solution. Arattai would have needed a different one. On product quality: the product gap is real but secondary — even a WhatsApp-feature-complete Arattai would have found no one to message.
0K+
Estimated week-1 installs
Based on #1 App Store rank
0%
Signal global download surge
Same window — won the narrative
0%
Arattai download collapse
Within 8 weeks of peak
Step 8 of 8
What I would have done differently
If I had been PM on Arattai and knew a controversy-driven surge was coming, here's what I'd have built — and what I'd explicitly not have done.
1Design the cold-start strategy before writing the first line of product code
The question 'how does the first user exchange their first message?' should have been answered before launch. Options: seed through existing Zoho Workplace (enterprise) users who could create family groups, a 'bring 5 contacts' referral unlock, or a geo-concentrated launch in one city to build critical mass locally. Any of these would have been better than nothing.
2Treat the controversy window as a retention sprint, not a marketing sprint
When the WhatsApp controversy created the surge, the product team should have gone into all-hands retention mode: build contact import in 72 hours, push 'your contact just joined' notifications, run group migration features. Every engineering hour during the controversy window should have been spent on activation and retention, not new features.
3Own the 'Indian messaging' positioning explicitly before Signal took 'privacy'
Signal won the global privacy narrative. Arattai couldn't compete on that. The differentiated position was 'built for India, by Indians' — local language support, India-specific features (UPI integration, regional language keyboards), Zoho's existing trust in Indian business. That's a position Signal can't occupy. Arattai never claimed it explicitly.
4What I would NOT have done: scaled marketing during the controversy peak
The worst possible use of the controversy window was pouring marketing spend into downloads. Every additional install without a working activation experience is a future uninstall. Marketing spend during the peak would have accelerated the vanity number and made the collapse look even worse in retrospect. The constraint was activation capacity, not download volume.