Welcome to My Blog

Late last week, it was discovered that Instagram had lowered the rate limit for its Platform API for some developers and completely removing access to it for others — without warning.

While these changes remain unconfirmed by Instagram, it appears that the visual content-sharing app has somewhat drastically decreased the frequency with which developers can use the Platform API to collect information from Instagram.

If that makes very little sense, don’t worry — you’re not alone.

But marketers should pay attention to this development and what it means. I had our own developers and engineers weigh in on how and why this happened, as well as the implications it could have for the rest of us. 

Our Experts Weigh In: Instagram Has Limited (or Cut Off) API Data Available to Developers

What’s Instagram Doing?

To get a better understanding of how API (which stands for application program interface) works and what these changes from Instagram mean, I turned to Sam Mallikarjunan, HubSpot’s Marketing Fellow, who used this analogy:

“Think of API as two parties, or two software systems, on opposite sides of the planet. The API is the cellular network that allows them to talk to each other. What Instagram has done here is lower the number of minutes included with your cellular plan.”

Here’s how that applies here: Prior to these changes, Instagram had a “rate limit” (the minutes analogy Mallikarjunan used) of 5,000 calls per hour, which are essentially requests for data on someone’s activity on the app — such as Likes and comments. On Friday, that rate limit was abruptly lowered to 200 calls per hour for some developers, and for others, access to the API was revoked completely.

According to Recode, very few, if any, developers were alerted to this change — and when I tried to access Instagram’s limits documentation, I was met with this result — suggesting that the app could be in the process of updating its written policies.

Source: Instagram

So, while Instagram has changed the frequency with which third parties, like developers, can ping the app to gather information, it hasn’t changed the type or amount of information that can be collected — unless, of course, that developer is part of the group whose access to the API Platform has been revoked completely. 

Which is interesting, considering that many believe this move was just the latest by Facebook — which owns Instagram — to put protective protocols in place in the wake of alleged misuse of personal user data by analytics and profiling firm Cambridge Analytica. It’s what Dmitry Shamis, HubSpot’s director of web development, calls “very reactive, instead of proactive.”

What Do These Changes Mean?

To further understand the impact of changes to Instagram’s Platform API, I once again turned to Mallikarjunan to explain how developers or other third parties might use it.

“Text messages also work as a metaphor, if anyone remembers a time before SMS was unlimited,” he said. Whereas one could text Instagram 5,000 times per hour to ask a question about someone using it, now, “I can text Instagram 200 times per hour.”

Of course, that’s only if these users have opted-in. Think of sites that give you the option of logging in with Facebook or Google, rather than using your email address and filling in several fields with information like your name, location, et cetera. Some sites also use Instagram in this way, perhaps not as frequently. 

As a developer, Mallikarjunan explained, “the things I might ask about are, ‘When is Amanda’s birthday?’. That’s one API call. Or, ‘Who has posted a photo with the hashtag #SamRocks?’ That would be a second API call.”

And as long as the users from which the developer is seeking this information has given permission for the third party to ask these questions, they’ll be answered. “But once I ask too many questions within a short period of time,” Mallikarjunan says, Instagram may cut you off. 

Source: Instagram

“For the standard everyday marketer, the impact won’t be huge,” says Shamis, since a volume of 5,000 calls per hour likely isn’t necessary for most. In fact, he explains, at that rate, “Something funky is probably going on.”

That may have been what was going on with Cambridge Analytica when personal user data was alleged misued, says HubSpot Senior Systems Engineer Ben Becker, when a professor and app developer introduced an app for people to install via Facebook, granting the app developer permission to access and use that data. What the developer wasn’t supposed to was transfer that information to Cambridge Analytica.

To get an idea of how a high API rate limit might be (ab)used in a situation like that one, Becker says to apply the numbers to that context. “Person A installs the professor’s app, giving the professor access to Person A’s information,” he explains. “At the time, Facebook allowed Person A’s information to also include data from his or her friends who Liked the Page or developer behind the app, which allowed the professor to also scrape their data.”

Now, Becker says, imagine a developer doing this to 50 million people — the estimated number of users for whom Cambridge Analytica obtained personal data — at 5,000 calls per hour. It could require less than 10,000 hours, or just over 417 days (less than a year and a half) to collect that amount of data.

Why Is Instagram Doing This, and What Does It Mean for Marketers?

“Right now, Facebook is in panic mode,” says Becker — which could explain the reactive nature of these changes, as Shamis described them. “I think this Facebook over-adjusting.”

But one has to wonder how good actors — as opposed to those pinging the API Platform for less-than-savory reasons — could be impacted by them. 

First, developers whose permissions were completely revoked likely faced that outcome because they were “probably bad actors, or  reasonably assumed to be a bad actor,” Becker says. “If you’re a good person, there’s nothing to worry about.”

But in terms of how the “good person” is impacted here, says Mallikarjunan, it might force them to be more selective about the data they want to gather using Instagram’s API Platform, seeing as they likely won’t be able to obtain as much info in a single timespan as they once were.

However, he suspects that could change, and wouldn’t be surprised if Instagram will allow higher rate limits to those who pay for them, perhaps “\under the pretense of privacy audits.”

So, why make limit these changes (for now) to Instagram, and not to Facebook at large? While Facebook did announce last week that it would be shutting down Partner Categories — the product that allows third-party data providers to supplement advertisers with targeting information directly through the social network — we haven’t heard anything yet about restrictions being made to its API access. 

Mallikarjunan believes it might have something to do with what is likely a significantly smaller subset of users who log into sites using Instagram — especially considering the limited number of sites that even offer it as a login option. (Personally speaking, I’ve barely ever come across it, but will certainly be keeping an eye out for it moving forward.)

Instagram has a “smaller user base,” he points out — 800 million monthly active users as of September 2017, compared to Facebook’s 2.13 billion, “and less data you can pull,” making it a convenient place to experiment with API restrictions.

Source: Statista

But as that suspected experiment progresses, we could see further restrictions made on both Instagram and Facebook, especially with the General Data Protection Regulation (GDPR) coming into force next month. 

But all of these events — starting with the alleged weaponization of Facebook by foreign agents to spread misinformation and influence the 2016 U.S. presidential election — and the resulting fallout from them is leading to pivotal moment for social networks, and the way they collect or leverage user data. 

And out of these social networks, “Facebook is the biggest, and it’s getting the brunt of the focus,” says Becker.

We’ll see how it plays out — and which other networks follow suit. 

Featured image credit: Instagram

If you work in tech and haven’t yet downloaded the app Blind — well, you’re missing out.

Blind calls itself “an anonymous community app for the workplace,” where employees of various tech companies gather to namelessly ask questions, leave feedback, or start discussions about everything ranging from salary to industry gossip.

So when Blind surveyed its user base to find out how many tech workers planned to delete their Facebook accounts — inquiring minds had to know what the results were.

What’s #DeleteFacebook?

The survey came in light of a recent scandal around Facebook, when it was revealed that analysis and profiling firm Cambridge Analytica may have abused personal user data, which allegedly should never have been in its possession in the first place.

Among other responses to the allegations — which included outrage, criticism, and the occasional lack of surprise among some — the hashtag #deletefacebook began trending on Twitter, reflecting the widespread sentiment and desire to delete the app and one’s account on it for good.

Blind decided to run a survey of its users to find out how much the #deletefacebook trend resonated among tech workers, collecting a total of over 2,600 responses between March 20 and March 24 of this year.

The Survey Results

Overall, 31% of respondents answered “yes,” whereas 69% answered “no.”

The responses were especially intriguing when broken down among Blind’s top five tech companies, which were determined based on the highest volume of responses. Facebook’s own employee responses were accounted for, with 2% of that base reporting it would delete the app.

It’s worth noting that the second-highest company responding “yes” to #deletefacebook is Snap Inc.: the parent company of Facebook’s rival app, Snapchat, which the former has been trying to emulate by way of new features and capabilities for many years.

That began after Facebook’s failed attempt in 2013 to acquire Snapchat — and it’s worth noting that Facebook also owns Instagram, which is arguably the most similar property to Snapchat.

Notice that ride-hailing app Uber ranks third — reminding me of the 2017 trend to #deleteuber that came after a series of PR crises for the company. One of them was a business move that many interpreted to be Uber’s capitalization on a U.S. executive order calling for a travel ban from certain countries.

Scores of people are deleting Uber after the company serviced rides at JFK airport while taxis were striking against muslim ban #deleteuber pic.twitter.com/D8cJMlxOxQ

— Josh Butler (@JoshButler)
January 29, 2017

It’s also interesting to note that 38% of respondents from Google said they would delete Facebook, seeing as the company has faced its own backlash and negative headlines pertaining to user data in recent weeks, as well. In fact, when Facebook CEO was called upon Congress to testify on these events, he was invited to do so by some committees alongside Google CEO Sundar Pichai and Twitter CEO Jack Dorsey.

Note that Twitter is absent from the above graphic — suggesting that response rates from that company weren’t high enough to include in the top five.

Blind has informed us that Amazon had the highest response rate, with employees from that company accounting for roughly 11% of the survey’s total base. Within that segment, 34% of Amazon employees said they would delete Facebook.

What Do These Survey Results Mean?

For those who don’t work in tech — just how seriously should these results be taken?

Not very, says HubSpot Product Marketing Manager Marcus Andrews.

“There is a precedent for this sort of thing. Facebook makes a change or does something that is received negatively, and there is a lot of backlash in the press and among individuals,” Andrews says. “I personally will not delete Facebook. It’s at the forefront of technology and innovation, and I afford them some slack.”

Henry Franco, HubSpot’s social campaign strategy associate, also encourages taking the survey results with a grain of salt, largely due to the audience surveyed. 

“This survey involved a tech-savvy audience that seeks anonymity so it can air its grievances about tech companies — an audience that might be predisposed to dislike Facebook,” Franco explains. “Tech workers are probably more likely to want to delete Facebook than the general population for similar reasons.”

It’s an interesting point. Similar to Andrews, I also chose not to delete Facebook, though I have deactivated on more than one occasion to take a break, prior to the Cambridge Analytica crisis.

HubSpot’s Co-founder and CTO Dharmesh Shah also recently shared his thoughts and insights on the matter.

“I know I’m likely in the minority of my tech circle friends, but I have no plans to delete Facebook from my life.”

– Dharmesh Shah

And while seeing my full Facebook data file — which may be in the possession of parties of which I’m not aware — was weird, I’m inclined to believe my decision not to delete my account is largely because I work in tech. After all, I spend my days reading, reporting, and analyzing news like this, which could be one reason I wasn’t entirely shocked by the Cambridge Analytica news (call me “jaded” or “cynical,” if you must).

Plus, I’ll return to my earlier note about Facebook’s ownership of Instagram — as well as Messenger and WhatsApp. As Andrews pointed out, the survey didn’t measure was how many of the people who said “yes” to deleting Facebook would also plan to delete its three umbrella apps.

For that reason, he says, “It’s just hard for me to take people seriously when they say they’ll delete Facebook.” 

But as Franco notes, we won’t know for sure just how much stock to put into these survey responses until Facebook itself releases data and metrics on its user numbers, which he says the company traditionally does on occasions like earnings calls.

“The fact is, there are a ton of reasons not to delete your account, as Facebook has a strong product with a deep moat,” he explains. “Doing so might ultimately be more of a burden to the user, considering their data are also being collected by a number of other tech companies.”

And when that earnings call takes place — we’ll let you know what we find out.