It’s Tough Being a Developer on the Messenger Platform
A few years ago, I listened to a podcast in which Margaret Gould Stewart VP of Product Design at Facebook, described how it took six months to redesign the Facebook “like” button.
As the founder of an early-stage company, I listened in jaw-dropping awe at how much time and effort went into one seemingly insignificant decision. Stewart made a strong case for it though and I could see why, at that scale of product usage, the paradigms are different.
At ServiceDock we work with the Facebook Messenger API and, for whatever reason, it seems like nowhere near the same level of due diligence is applied to decisions relating to business messaging.
The following points cover my major bugbears but I’m sure many other Facebook Developer Partners (DPs) face different issues.
1. Net Promoter Score Question
Even if you don’t know what Net Promoter Score (NPS) is you’re likely to have provided an answer to an NPS question at some point in your life (we explain exactly what it is and how to calculate it here). It is the gold standard question that businesses use to determine customer satisfaction and advocacy levels. You will often be asked an NPS question after an interaction with a customer service agent for example.
Businesses put enormous effort into tracking and improving NPS over time because it is considered to be an indicator of future business success (ie if customers are willing to advocate on your behalf you will get more customers through word-of-mouth marketing So your business will organically grow). In many cases, staff incentives are in some way tied to NPS results, so any errors in calculating it can materially impact the take-home pay of employees. This is serious stuff.
The question follows a standard format and the answer option must be a scale from zero to ten. This means the customer must have a choice of 11 digits and that is almost certainly the reason a maximum of 11 options are permitted for quick replies on Messenger. Originally it was limited to 5 replies (might have been 7) but after lobbying by us, and presumably many other organisations, producing this version of the question became possible on Messenger at some point in 2017:
But when Messenger 4.0 landed at the end of 2018, the Messenger UI only facilitated showing the NPS question like this:
What is the problem?
The problem with this change is many consumers do not know that swiping left will reveal more answer options or if they do, they make a mistake and hit the number 4 option when trying to swipe. Our data puts the error rate at around 6%. That might not sound huge but anyone familiar with how NPS works will understand that a 6% decrease in promoters (assuming all errors wanted to hit 9 or 10) and an equal increase in detractors (all hitting 6 or less) will result in a massive 12 point slip in NPS.
This error rate means that customer interactions on the Messenger platform will show a lower NPS than interactions on another channel simply because of the way the question is formatted. So even if consumers rated the Messenger interaction 5 points better than a similar email interaction for example, the error rate would make it appear that customers prefer customer support via email rather than messaging.
Messing around with NPS in this way is equivalent to Messenger shooting itself in the foot. We have already seen cases where businesses have dropped Messenger as a channel simply because the NPS Score was markedly lower than comparable channels such as webchat for example.
For us, it’s infuriating beyond belief and shows a real lack of commercial sensitivity on behalf of the Messenger team. One basic principle of good design is that form follows function but that principle seems to have been thrown out the window here.
2. Messenger Code Deprecation
In April 2019, DPs were advised that the Messenger app would no longer be able to scan Messenger Codes from August 15th 2019. Again, for anyone unfamiliar with what a Messenger Code is, it is basically a QR code that can only be scanned with the Messenger camera.
We have been using this to help our retail and restaurant chain clients engage consumers in their stores and restaurants. We have effectively done a proof of concept that has proven that this works really well and is much more effective than QR codes in certain scenarios.
The reasons given by the Messenger crew for the decision to deprecate the Messenger Code were:
- They “believe [accessing Messenger pages by scanning codes] will be more seamless by leveraging the phone’s native capabilities to scan QR codes with m.me links.”
This would be a much more reasonable statement if every man, woman and child on the planet had an iPhone, which can now scan QR codes with the native camera application. But the majority of phones in circulation out there are Android phones and there is a myriad of different ways to scan a code depending on whether it’s a Pixel phone, Samsung or another brand and even models within the Samsung range. My field tests indicate that your average Android user does not know how to scan a QR code with a native feature on their phone and few are bothered to download a QR code app.
2. They weren’t being heavily used by consumers.
This, of course, is primarily because consumers didn’t know the feature existed. I’m not going to explain that in any detail but the below illustration shows how users were expected to discover Messenger Codes and how difficult that was to explain concisely. It also shows a couple of very straightforward alternatives that might have helped with adoption. I would love to know what sort of user testing went on when this was implemented.
What’s the big deal anyway?
There are other reasons why Messenger Codes will be a huge loss to us and the Messenger Platform in general.
- Trust — Consumers can trust that a Messenger Code is taking them to Messenger. A QR code can be taking them anywhere.
- Novelty — Consumers are largely blind to QR codes (in the Western World) and let’s face it they’re ugly. Whereas the Messenger Code is something new and it’s much more attractive.
- Messenger Branding — Messenger Codes present a great way to brand Messenger as a channel in the physical world.
3. No Customer Support
For all the riches that Facebook has and for all the potential that DPs bring to the platform, for some reason Facebook Messenger’s management team can’t find it in their hearts to provide a proper customer support desk for us. There are some great community resources out there like the Messenger Platform Developer Community run by Lars Schwarz but very often even Lars isn’t able to help DPs with their issues. In many cases conversations in the group turn into group therapy sessions of exasperation sharing.
I’m not going to say any more than that but seriously, with $13.2 billion in revenue, you would think they could spare a few hundred thousand to help out their DP community. It might have helped avoid some of the scandals that have happened, as honest DPs would have been asking questions about things that would have helped identify loopholes that dishonest DPs were exploiting.
Conclusion
“Move fast and break things” helped Facebook grow like a rocket ship and maybe the Messenger Team still lives by that mantra. But either Facebook wants real businesses (i.e. DPs and our clients) on the Messenger Platform or they don’t. Assuming the strategy is the former, then better lines of communication with DPs need to be developed.
My best guess as to what’s happening here is that a group of extremely affluent California based technophiles are making decisions based on their own experiences and environment. I’m happy to be corrected, but there’s an apparent absence of real-world people who know how terrible the QR code scanning capabilities of many Android phones are and that something like Google Lens (that does scan QR codes) is not on every phone and is not used when it is. Or maybe they haven’t grasped the fact that features need to be easy to discover in order to build traction or could it be that they simply don’t appreciate the huge problem that moving offline consumers to online continues to be? Understanding the importance of NPS scoring in commercial businesses and in rating different channels also seems lacking because what they’ve done makes Messenger look like it underperforms versus older support channels. And most tellingly, while they’ve been very successful at building an enormous DP community, they seem reluctant to engage directly with us in two-way conversations. Taking that approach means things are likely to get worse instead of better.
Ahh… I feel a little better after getting all of that off my chest. I could have done a whole blog post on the authorization process but I ran out of steam. However, if you want to vent about working with the Messenger Platform, please feel free to vent in the comments below.
