I have spent the last few months in Bikaner, a Tier 2 city in Rajasthan. My extended stay offered a unique opportunity to observe how users interacted with apps on their mobile phones – I could watch them from a distance, even as I was embedded in their context.
My small sample size comprised about two dozen users and my interaction with them consisted of informal chats, not a programmed study. However I did experience the ‘saturation of knowledge’ described by sociologist Daniel Bertaux. According to Bertaux, “the researcher is surprised or learns a great deal from the first few interviews. By (say) the fifteenth interview, the researcher recognises patterns in the interviewees’ experiences. More interviews only confirm what the researcher has already sensed.”
To validate my findings, I also shared them on this tweet thread and a number of people confirmed that they had witnessed similar patterns. I hope my observations will prove useful for people who are building for the next billion users and help them to create more representative user personas.
The Importance of Placement
The users I met constantly share content from YouTube, WhatsApp and earlier, TikTok. However, as I watched them share content, I realised they were not using the share icon at all – in fact they couldn’t even identify it. When I asked them how they could share so many videos without knowing what the share icon was, this was their reply: “We know where to tap on the screen. Then the WhatsApp icon pops up and we are in familiar territory.”
On further probing, I discovered that many of these young users had discovered video sharing via Chinese porn apps. The Mandarin interface was incomprehensible to them, but they learnt to navigate by becoming familiar with the position of the icon on the screen. As a result they associated the functionality with the location of the button on the screen, not with the icon or caption. So if you are contemplating changing the position or orientation of buttons or icons in your app, think again.
Minding the Parents
The dominance of joint families in small towns has a significant impact on user behaviour. Young people may pay from their own pockets for a movie or a snack at a tea stall, but they will invariably opt for the ‘Cash on Delivery’ (COD) option on e-commerce purchases. Most doting mothers will mask these purchases under the larger household budget and not report them to the father. As a product designer, my take away from this observation was that if you are running a promotion targeted at young people in Tier 2 cities, ensure that you have a COD option.
While it may seem counter-intuitive, living in close proximity to parents is also the reason that young people prefer voice messages to text chats on Whatsapp. Parents are wont to spy on their children’s phones and will mistake the voice chats for songs. (We don’t know how long it will be before they catch on and start keeping tabs on voice messages!)
Speed Constraints & Delays
I compared the page load speed of popular apps like Paytm, Google and Facebook on my phone vs the other users’ phones. For a device with similar processing power under the same network conditions, the page loaded twice as fast on my phone. Most devices are infected with malware, which people have installed unknowingly. This blocks CPU cycles and bandwidth while syncing. Let’s also not forget the storage-guzzling good morning messages from WhatsApp. The result? Your carefully designed, legit lightweight app becomes very slow.
One Time Passwords (OTP) are often delayed and can take up to two minutes or more to be delivered. Users have figured out that if you reboot the device, the OTP gets delivered faster. This hack is fine if you are making a payment on a desktop, but if you reboot a mobile device in the middle of a session, the transaction is lost.
There could be multiple solutions to this, which will need to be tested at scale before adoption. For example, send the OTP as a Class 0 pop-up message or a USSD message. Even simpler would be to increase the interval for the request to resend the OTP or if feasible, not disabling the previous OTP while sending a new one.
Buying Online
Most people associated e-commerce with Flipkart and then Amazon. E-commerce purchases are collaborative – kids help with navigation, usage and discovery, while parents step in for payment authentication and pricing approval.
Paths to promotions vary
People discover promotions through their friends or promotion-specific WhatsApp groups. Therefore, when making a purchase, most users will not look for promotions in between transactions.
Payments like electricity or recharge are done by parents using their kids’ accounts and they will check the promotions section before making a payment.
Voice searches need to improve
People hate typing: They are happy with the voice to text feature and voice messages on WhatsApp. When it comes to voice searches in e-commerce, however, they say the results are not good. I wonder if anyone has analysed the Click-through-rates (CTR) of search results to compare voice vs type input.
Imported vs Made in India
When it comes to promoting an unbranded, non-consumable product in a low-mid price range, words like ‘Imported’ or ‘Export Quality’ are effective as they are seen as proxies for product quality. For consumable products ‘Made in India’ works best, perhaps because it is seen as a proxy for freshness.
Lingerie & sleepwear are DAU not ARPU
If you are running a horizontal marketplace and you are not happy with your ‘Page views’ divided by ‘Add to cart’ conversion statistics, here’s something you should try: Remove page views contributed by the lingerie and sleepwear category. It is entirely possible that your daily active users on those pages are young males who have no intention of making a purchase.
The Future is QR
QR codes are under-hyped. The users I met were not comfortable typing URLs, but they were fine with scanning QR codes from promotional pamphlets. From shops to tuition services, beauty salons to pharmacies, the preferred advertising route is to distribute pamphlets with QR codes along with the newspaper. People are accustomed to saving numbers by scanning codes. In fact I saw young people adding each other as contacts on WhatsApp, not by sending a contact card, but by scanning a QR code.
“The entry point for PC internet is the search box. The entry point for mobile internet is the QR code.”
Alan Zhang, Founder, WeChat
Paytm: All-purpose QR Code reader
It is widely known that you can use Paytm to scan all QR Codes, including Non-UPI / Non-Payment QR codes. People also know that Phone Pe and Google Pay do not offer this facility.
Right now, Paytm QR scan serves as a conduit for a QR code destination. In the near future, it could potentially become the smartest way to divert traffic to other Paytm services. Given how well it has captured other use cases, it is surprising that WhatsApp has let this go.
Galla-Bahi-Tally
In my research with over two dozen shop keepers, I found that their most preferred option was cash, followed by e-payments. POS swipe machines were the least preferred. This was especially true for ticket sizes above Rs. 2000. Most people I spoke with tended to use Google Pay, because of high success rates and bank-to bank integration. The one player that seemed to be missing in action was Phone Pe.
Even though my base was small, no one used – or wanted to use– any of the kirana digitisation apps. This is because their workflow is different and consists of ‘Galla-Bahi-Tally’
Galla: The cash maintenance and handover when there is a change of shift between the owners, typically a father and son
Bahi: The record of debt given to customers and payment made to suppliers
Tally: The reported income, usually a fraction of the real figure
Merchants have always kept these three separate and as far as I could tell, it will be difficult to convince them otherwise.
The Sleeping Giant
The amount of order processing that is being done on WhatsApp is mind-blowing. What’s more, it is shocking how broken the whole experience is. There is no convenient way to share or browse catalogs, no inventory management or cart preparation.
Yet despite the tedious back and forth, people persist because of Whatsapp’s sheer ubiquity and universally-understood interface. My takeaway is that Whatsapp is a sleeping #dukaantech #kiranatech giant, which will prove life-threatening for other apps, if awakened.
Most observations make sense except the absence of Phone Pe part. The ‘why of it’ isn’t there. Given Phone Pe has a far better UI than G-pay what could be the likely reasons for the same?
Agree. I was intrigued by that too.
Some of these are familiar and validate existing understanding and some are fascinating – eg. understanding icon placement from Chinese Porn Apps – I’d be curious to know how this conversation would have gone, because people don’t always know where they get their habits from.
Interesting lesser knows observations! Thanks for sharing.
Fantastic observations! Thank you for sharing this!
Observation is quite good but i think one point is missing related to “language”.
Do they use “hindi” as a default language in their phone ?
Some really fascinating observations – thank you for sharing and hope to see more articles like this!
As a startup founder, I can tell you that the big issue is that researchers tend to listen to what respondents tell them. Especially in the case of joint families this will be false- kids never want to admit they have no control and will give all kinds of answers. As Prashant says, you have to be embedded in their context to know what is unarticulated. Same with merchants – they will say Yes Yes, just to get you out of their shop and get back to dhanda.
Listening is not the only research methodology. Users don’t always say what the problem is so choosing the right methodology is very important to gain the right insights.
I hv a qs for Prashant – if you were VP Product at Paytm, didn’t you have these learnings when you were working there? Surely Paytm will have had a huge Bharat-research team? Is it that you can only share now, because you are no longer working there? This is a genuine qs, not trolling. I just want to understand what is the level of user research in these mega cos. TIA.
No . We had a Research team but this information was mostly gathered after I left Paytm .
Fantastic story – we need one place to capture all this info. People sitting in ac offices on iPhones cannot design for Bharat. All it takes is the will to interact on ground with users who are not like you
Really on-point observations. Thanks for sharing Prashant. Maybe this knowledge exists inside big cos – but it is not avlble to all of us
If the observation is that people understand more in the voice message as compared to scrolling text…. why don’t u implement that here as well… y so much to read then…4min read time heads up definitely help but then when u look up the article on a small mobile phone screen it definitely does not look like 4 min thing and then readers loose interest!
Thanks Srishti, we agree that long reads are hard on a mobile. We’re trying to figure out a solution to this. Think that just having voice instead of text may not be the solution, because you can scan thru text quickly but voice is linear and you have to listen to the whole thing. Working on it 🙂