How Easy is it to Rig WeChat? Super Easy ... and Super Cheap
Like most everyone in Beijing (and by extension, all of China), I've been spending a lot of time on WeChat these days.
And as someone who creates content as well as consumes it, I've often looked at the fantastic view counts of some of the stuff being circulated around WeChat and walked away dutifully impressed – 20,000 views, 40,000 views, even 100,000 views in some Chinese language accounts.
And with the Beijinger account typically getting views far less than that (our best post ever received 7,900 views), I've been despondent about our low view counts.
Is our stuff really that unattractive?
OK, so maybe we send a dud or two once in a while, but we keep it limited to stuff we think our readers will be interested in (and the occasional commercial message that keeps us in business). Could we really be missing the mark by a factor of 10 or 100?
A clue: everyone cheats
Last year a staffer joined our company from an agency that specialized in WeChat marketing told me the reality: WeChat view counts (and likes, and votes) are easily manipulated, and that in her estimate, 95 percent of the accounts regularly pay to inflate their view counts.
Let that sink in for a moment: 95 percent of the publicly stated view counts, likes, etc are faked.
This was news to me, as I was under the impression that WeChat, because it was connected to individual telephone numbers, was impossible to rig: one phone user, one view. Sure, maybe some fool with a few dozen phones could manipulate a view count by a few dozen, but tens of thousands of fake views? Not possible.
How naive I was.
Our policy: Let counts speak for themselves
We've always followed a policy of open view counts and organic growth on our social media accounts (as well as this website). There have been times in the past where we've paid for advertising on Facebook to get additional traffic or encouraged people to follow our social media accounts via incentives – but we've found that for the most part, these are great ways to draw short-term attention to your accounts but not a sound strategy for long-term success.
Though we've never participated in such shenanigans, I've been endlessly curious how it actually worked.
The Test: Rig an internal poll
So last week we put it to the test and discovered exactly how easy (and cheap) it was.
What sparked it was an internal staffers' suggestion that we do our annual restaurant awards via WeChat vote. I didn't like the idea as I'd seen some suspicious-looking results from polls on WeChat – one magazine's recent WeChat-based Best Restaurant awards I've been observing were turning up some unusual results.
As someone who has paid close attention to Beijing's restaurant scene for close to 15 years, I'm fairly sure I know which venues are popular amongst expats. And let's cut to the chase: while each of us plying the expat rag trade likes to think of our audience as entirely unique, there's enough overlap between the audiences to assure the results are generally in line with one another.
And when a restaurant that never appears even in our nominations ends up winning a WeChat vote, there's usually something fishy afoot.
So just to show check for ourselves how easy it is to "game the system" on WeChat, we decided to rig a poll of our own.
After a cursory search via Taobao, dozens upon dozens of people offering such services were found, with ads like the one below:
After a few minutes of haggling and a transfer of a small amount of cash, we "bought" 15,000 votes in a fake contest we created. We chose 15,000 votes (the cost, by the way, was 1 kuai for 10 votes) because we knew it was an impossible figure for us to achieve, given it is more than twice as high as the most popular post ever in our year-plus history of using WeChat.
Thus we created this poll:
The poll was not even circulated publicly – we merely sent it to ourselves as a "test" post. You can see a few of the votes cast by our staff in the picture at the above right.
We told our chosen service provider to choose the 15,000 option in the poll. Ten minutes later, they delivered:
So how did they do it? Did they pay off 15,000 people to vote in our poll? No, they simply used software that hacks the results with almost no effort on the part of the seller. Page views and likes are easily purchased too, for even less (we decided not to pay to inflate those numbers, hence only 18 views).
One Beijing publication currently holding WeChat votes for the city's best restaurants is seeing winning restaurants garnering somewhere in the vicinity of 100 votes, sometimes less. That means it doesn't take much – say, an investment of RMB 20 kuai for 200 votes – to sway the vote. If I were running a restaurant, I'd be mighty tempted to spend the equivalent of a taxi ride to secure my victory.
This makes awards results descend into simply a matter of who's willing to pay the most to win. And it's dirt cheap.
That unfairly robs hardworking restaurants who don't "pay to play" with a chance at glory, and it indirectly calls into question everything we do, as many people tend to lump us expat entities together, despite the fact that we're all quite different and have different standards of ethics and fairness.
Whither WeChat polls?
That's why we've decided not to conduct polls by WeChat at this time – there's just no veracity. We spend an inordinate amount of time investing in keeping our reader polls as honest as we can make them, and it's just too easy for WeChat as a system to be gamed.
We prefer to stick to the way we do it now: via a web-based poll that requires more than a reflexive click of a button (or a 20 kuai payment) to respond.
But, you may ask, isn't it true that people can pay to have fake ballots entered in our polls just like they do with WeChat? In short, yes ... and in fact some do.
But when you design your polls the way we do, it's easy to detect such vote fraud. Our polls typically require the voter to answer a variety of demographic and logical questions that can easily snag someone who is stuffing the ballot box.
In additon, our survey software captures extensive data on each voter that allows us to detect and eliminate voting fraud – far easier than it is to detect vote fraud on WeChat.
What about people who offer incentives for real people to vote?
But what if the venue gets all their friends and their customers to vote, guaranteeing that only the most popular venues win? Yeah, that happens. Welcome to democracy in all its glory (and faults).
But I'd certainly trust the quality of a place that wins on the backs of encouraging their customers to vote over one that spends RMB 20 to purchase victory.
So what are we to believe about WeChat read counts, vote tallies and likes?
In short: if numbers look too good to be true, it's because they are.
We have about 15,000 followers in our WeChat account. What we'd consider a "hot" post would get over 2,000 views, which is about 13 percent of the subscribers – and is in line with typical open rates for our email newsletter. Interestingly, that's also roughly in line with view counts on our blog as a percentage of overall visitors to the site – divide blog views by unique visitors and you get about that percentage.
WeChat: Just another source of (overwhelming) information
So it's important to look at WeChat for what it is: just another platform, and one that is quickly becoming overwhelmed with information.
At one point WeChat was a great new channel with lots of signal and little noise – nowadays its simply another fire hose of information, overcrowding your mind along with email, Facebook, Twitter, Weibo, Instagram ... you name it.
WeChat in fact is much like email newsletters: at the beginning, receiving an email newsletter seemed so direct and intimate and targeted and special; but over time as you gather more newsletters, your email box becomes cluttered with more information than you can absorb, so you practice a natural triage and end up deleting most of them without reading.
When I first started using WeChat I found it refreshing, and I was able to keep up with it – I could easily cruise through my contacts' Moments and finish reading the material sent by the few official accounts I follow in a few minutes.
Now I delete most things before looking and only cursorily check my Moments, and I find myself blocking the Moments of most.
WeChat is here to stay (until the Next Big Thing comes along to replace it).