Our community continues to grow, now with nearly 2.1 billion people using Facebook every month and nearly 1.4 billion people using it daily. Instagram also hit a big milestone this quarter, now with 500 million daily actives.
The reason I’m talking about this on our earnings call is that I’ve directed our teams to invest so much in security on top of the other investments we’re making that it will significantly impact our profitability going forward, and I wanted our investors to hear that directly from me. I believe this will make our society stronger, and in doing so will be good for all of us over the long term. But I want to be clear about what our priority is. Protecting our community is more important than maximizing our profits.
Over the next three years, the biggest trend in our products will be the growth of video. This goes both for sharing, where we’ve seen Stories in Instagram and Status in WhatsApp grow very quickly, each with more than 300 million daily actives, and also for consuming video content.
In messaging, today already more than 20 million businesses are communicating with customers through Messenger. Now we’re starting to test business features that make it easier for people to make the same kinds of connections with businesses through WhatsApp.
We’re now using machine learning in most of our integrity work to keep our community safe. When Hurricane Maria hit Puerto Rico, we used AI to look at satellite imagery and identify where people might live and need connectivity and other resources. Progress in AI can unlock a lot of opportunities.
Facebook has over 6 million active advertisers, and we recently announced that Instagram has over 2 million advertisers. The vast majority of these are small and medium-sized businesses, which are a major source of innovation and create more than half of all new jobs globally. These businesses often have small ad budgets, so the ability to reach people more effectively is really valuable to them.
Video is exploding, and mobile video advertising is a big opportunity. Until recently, ads were only eligible for Ad Breaks if they also ran in News Feed. But in Q3, we gave advertisers the option to run ads in videos alone. We’re seeing good early results, with more than 70% of Ad Breaks up to 15 seconds in length on Facebook and Audience Network viewed to completion, most with the sound on.
I would say not all time spent is created equal. That’s why I tried to stress up front that time spent is not a goal by itself here. What we really want to go for is time well spent. And what the research that we found shows is that when you’re actually engaging with people and having meaningful connections, that’s time well spent, and that’s the thing that we want to focus on.
I do think your point is right that not all kinds of content can be supported by ads, no matter how effective we make that. That said, the current model that we have for at least getting some of the lighthouse content onto the platform is to pay up front. And what we would like to transition that more to over time and what an increasing amount of the content is, is revenue shares for ads shown in the videos. And as we do better and better on the monetization there, that will support people with higher production costs and doing more premium production and bringing their content to the platform. And we’ve certainly found on the Internet and YouTube and in other places that there are whole industries around creators with different cost structures than traditional Hollywood folks who can produce very informative and engaging content that a lot of people like and enjoy and that builds communities and that helps people connect together in a way that definitely can be supported by this ad model.
Turning to Services. Revenue reached an all-time quarterly record of $8.5 billion in the September quarter. A few quarters ago, we established a goal of doubling our fiscal 2016 services revenue of $24 billion by the year 2020, and we are well on our way to meeting that goal. In fiscal 2017, we reached $30 billion, making our Services business already the size of a Fortune 100 company.
The reason I’m so excited about AR is I view that it amplifies human performance instead of isolates humans. And so as you know, it’s the mix of the virtual and the physical world and so it should be a help for humanity, not an isolation kind of thing for humanity…Apple is the only company that could have brought this because it requires both hardware and software integration, and it requires sort of making a lot of – or giving the operating system update to many people at once. And the software team worked really hard to make that go back several versions of iPhone so that we sort of have hundreds of millions of enabled devices overnight.
But in terms of price elasticity, I think it’s important to remember that a large number of people pay for the phone by month. And so if you were to go out on just the U.S., since that tends to be more of the focus of this call, you look at the U.S. carriers, I think you would find you could buy an iPhone X for $33 a month. And so if you think about that, that’s a few coffees a week. It’s let’s say less than a coffee a day at one of these nice coffee places…In terms of the way we price, we price to sort of the value that we’re providing. We’re not trying to charge the highest price we could get or anything like that. We’re just trying to price it for what we’re delivering.
In the 18 years since Alibaba was founded, China’s per capita GDP grew by a compounded annual rate of 14%. By comparison, the per capita GDP of the United States grew 3% during the same period. We all understand the magic of compounding. When you compound at 14% rate over 18 years, which is the life of Alibaba, the average Chinese citizen is 10 times better off today than in 1999 with per capita GDP growing from $870 to $8,100.
Today, China’s per capita GDP is still only 1/7 of the per capita GDP of the United States. Based on the track record of sustained income growth over the past years as well as on the backbone of a modern Internet infrastructure and productivity gains from technology, I’m very optimistic that China will continue to experience real income growth for years to come. This will translate into a rising middle class characterized by ever-increasing and higher-quality consumption. And this long-term secular trend bodes well for Alibaba.
Our cloud computing business continues to defy gravity. Revenue increased by 99% year-over-year. We continue to multiply our product portfolio, including the introduction of a new relational database and a state-of-the-art server developed in-house that serve the needs of large enterprise customers.
Mobile MAUs on our China retail marketplaces reached 549 million in September, an increase of 20 million over June quarter. Annual active consumers on our China retail marketplace reached 488 million, a net add of 22 million from the 12 months period ended June.
And the key thing is that the data-driven logistic network, actually we are – Cainiao is not going to be a logistic company and we are not interested into building another logistics company. Instead, we will work with a lot of logistic companies, delivery companies to build a network across the world.
Our concerts business is our flywheel, attracting over 30 million fans to shows globally in the quarter, which then drove record results in our onsite ticketing and advertising business. Through October, we have sold 80 million tickets for concerts in 2017, up 20% year-on-year. Digging deeper into concerts, strong global demand for concerts through the third quarter drove a 16% increase in attendance to 65 million of fans at our 20,000 shows in 40 countries.
With the success of the concert flywheel, we’re promoting more shows for more fans, more effectively pricing and selling tickets and delivering a better experience than ever. As a result, we will spend over $5 billion producing concerts this year, making Live Nation far and away the largest financial partner to musicians.
With over 1,000 sponsors across our onsite and online platform, Live Nation is a global leader in music sponsorship, providing brands with opportunities to reach our core audience.
There is no artist that’s dying to put tickets on a secondary platform as a solution. That isn’t how they build their brands with their fans. What they want to do is figure out how to price it right and then make sure their fans actually have a shot to buy the ticket, not deliver it to the on-sale, have the scalper buy it, and their fan in Boston ends up paying 3 times the price.
We think Fan Verified and then you add on presence from the digital perspective are a real important combination for these artists of the future, who now believe they have some shot at controlling and delivering to their fans the price point at the exact price they want. And so, we think this is a pivotal product – suite of products that we’ve developed in Ticketmaster. It’s under a new division within Ticketmaster called Artist Master where we have a new leadership team waking up every day, making sure that we can deliver artist products, so the artists can deliver their tickets to their fans at better pricing and at the price they want.
In fact, there’s 10,000 unique parts, so to be more accurate, there’re tens of thousands of processes necessary to produce the car. We will move as fast as the least competent and least lucky elements of that mixture. So while the vast majority are going incredibly well, there are some problem areas.
The primary production constraint really, by far, is in battery module assembly. So a little bit of a deep dive on that. There are four zones to module manufacturing it goes through four major production zones. The zones three and four are in good shape, zones one and two are not. Zone two in particular, we had a subcontractor, a systems integration subcontractor, that unfortunately really dropped the ball, and we did not realize the degree to which the ball was dropped until quite recently, and this is a very complex manufacturing area. We had to rewrite all of the software from scratch, and redo many of the mechanical and electrical elements of zone two of module production.
The ramp curve is a step exponential, so it means like as you alleviate a constraint, the production suddenly jumps to a much higher number. And so, although it looks a little staggered if you sort of zoom out, that production ramp is exponential with week over week increases.
There’s vastly more automation with Model 3. Now the tricky thing is that when one automation doesn’t work, it’s really harder to make up for it with men and labor. So with S or X, because a lot less that was automated, we could scale up labor hours and achieve a high level of production. With Model 3, it tends to be either the machine works or it doesn’t or it’s limping along and we get short quite severely on output.
I think that we will be able to achieve full autonomy with the current hardware. The question is, it’s not just full autonomy, but full autonomy with what level of reliability, and what will be acceptable to regulators. But I feel quite confident that we can achieve human level – approximately human level autonomy with the current computing hardware. Now regulators may require some significant margin above human capability in order for a full autonomy to be engaged. They may say, it needs to be 50% safer, 100% safer, 1000% safer, I don’t know. I’m not sure they know either.
We are very excited about the increased momentum in 5G around the world. We are leading the industry and are accelerating the commercial launch of 5G across millimeter wave and sub-6 gigahertz in early 2019. We recently announced the world’s first 5G data connection achieved on the Snapdragon X50 modem chip set, and our leading 5G 3GPP standards development, ongoing prototype efforts and are supporting global 5G new radio trials.
Gigabit LTE is the first step in network operator’s transition to 5G, and there are now 41 operators in 24 countries supporting Gigabit LTE. We have demonstrated download speeds of greater than 1 gigabits using our X20 LTE modem in the U.S. with both Ericsson and Verizon, as well as Nokia and T-Mobile. Most leading device-makers are rapidly adopting Gigabit LTE into their device portfolios.
In the premium tier, our gigabit-enabled Snapdragon 835 now has more than 120 designs launched and in development, including recent flagship devices like the Samsung Galaxy Note 8, Pixel 2 and Pixel 2 XL, LG V30, and the Xiaomi Mi MIX 2. We have also introduced new high-tier and mid-tier Snapdragon products to further expand our competitive position in China across all tiers.
Lifetime value of a pet grew to $701 in the quarter, reflecting improvements in retention and in optimizing our cost-plus pricing strategy by subcategory. As lifetime value increases, we are able to increase our allowable tax spend and invest more aggressively in testing of new acquisition initiatives, with a goal of understanding whether they can meet our internal rate of return targets over time.
Total enrolled subscription pets increased 15% year-over-year to approximately 359,000 pets as of September 30. Monthly average revenue per pet for the quarter was $52.95, an increase of 9% year-over-year. Average monthly retention was 98.61%, consistent with the prior year period.
We want to be cash flow positive. The other guardrails are internal rate of return. And for any significant investments that we’re making, we are targeting internal rates of return of greater than 30% and hopefully, closer to 40%. When we’re doing some different testing or some long-term initiatives, maybe they’re a little bit lower, but on a blended basis, we want to be over 30%, internal rate of return closer to 40%. So those are probably the guardrails that we’re using. When I talk about how much more we’ve learned, it’s really driven around increasing the velocity of hospitals in our same-store sales initiatives. And order of magnitude may be we’ll invest an additional $2 million or $3 million next year than we otherwise would have anticipated. So definitely inside of our guardrails, but long term and foundational.
Well, the biggest opportunity, this is the overall market. I mean, we’re still in a very low underpenetrated market. And I’ll be saying this for the next 5 years. If you look at the number of pets that have some form of medical insurance, we’re over 1%, but we’re still under 2%. We want to try to grow the category growth. When we think about how to do that, it’s all around veterinarian first and making it normal and key. We have historically been growing the company by adding stores. Over the next few years, we feel it will become important for us to work on the same-store sales.
So as a reminder, our goal is to eliminate a slow cumbersome reimbursement model, where a pet owner fills out a bunch of paperwork and puts it in the mailbox and wait for 2 to 3 weeks to see if they are going to get paid. We think the problem we’re solving for pet owners is making it easier for them to budget for if and when their pet become sick or injured. And part of solving that problem is being able to pay hospitals directly at the time of invoice. Part of that solution is being able to integrate with practice management software in a way that makes the process very quickly. Our goal is to pay these invoices in under 5 minutes from the time those created. We like the results that we’re getting. But we want to speed up the velocity of the – number of places that we’re rolling it out now that we’re better at it. And we’re still in relatively early days to learn how to speed up the velocity, now that we like the results we get on a per hospital basis.
The General Data Protection Regulation, or GDPR is a piece of legislation passed by the European Union to enforce tougher more consistent guidelines on organizations that operate across the region and particularly, global Internet companies and deals with customer data and product deployments. These guidelines impact the millions of domain names that we have registered to European end-users. So, we need to comply along with everyone else by May 25, 2018. This will require a significant investment in engineering. However, its work that needs to be done by everyone across our industry and across many industries and it plays to our engineering strengths. Also, other governments are starting to the EU’s lead and pass their own legislation. So, while the work will cause some pain in the short-term if we do it well better and faster than others, it could create some opportunities in the future.
This deal gives us access to the Otono eSIM platform. eSIM allows mobile users to choose and switch between networks without needing to insert and activate traditional SIM cards. In coming quarters, eSIM should allow Ting to support high-profile smartwatches and other exciting new connected devices. In coming years, eSIM should allow Ting customers to move seamlessly between networks on almost any device imaginable.
In just four years the results have been dramatic, which Chegg now serving nearly ten million visitors in a month according to ComScore and we expect to have around four million paying customers in 2017. Further, we expect to have more Chegg Services subscribers than textbook customers for the first time in our history. Our Chegg Services revenue has grown from just $25 million in 2012 to what we expect will be over $180 million in 2017.
And given our new adjusted EBITDA guidance today, we will go from a company who was losing $16 million in adjusted EBITDA right before our IPO, to a company that is now forecasting nearly $45 million in adjusted EBITDA for 2017. This turnaround has helped us go from a company using over $100 million in cash each year to support a slow growth textbook business to a high growth, high margin learning company that now produces free cash flow. It’s been a successful transition and we feel like we’re just getting started.
As powerful as Chegg Study is already, we believe the service will become even more relevant to an increasing TAM, as we continue to expand its content and capability. To that end, we recently announced the acquisition of Cogeon, developer of the app Math 42, an adaptive A.I.-driven Math application, which has been downloaded over 2 million time. Math is the universal need and unfortunately a universal problem as 64% of U.S. students are not prepared for college level math, and over 40% of U.S. students take at least one remedial math course.
Even though students look to A.I.-driven tools to help augment their learning, we know that one-on-one human help will continue to be critical in the learning process, which is why we continue to invest in Chegg Tutors. In the third quarter, we saw the time students spent on the site increase, with an average student now spending 188 minutes in tutoring sessions throughout the semester in key subjects like computer science, calculus, statistics, finance and accounting.
The magic of what we are building is the ability to have access to millions of millions of millions of students and their records and their history, all the data about them it’s in our proprietary student graph, and then the ability to do matching based on that uniquely done by how we are capable of doing it versus others. So, we met with all the players in the space. Obviously, everybody would like access to this Chegg audience, but we believe what working on is going to be special and unique and has the opportunity to be huge.
So, we don’t see a reason with such high gross profit margins to change the price. But we do know we have significant pricing power, given the fact of not only the test results which you can just look at the usage. We have record usage in terms of how often they use it, how many pages they consume, what they use for the number of questions that they’re asking, the number of the subjects that they’re asking it in. I mean Chegg Study has become a beast, and we just — what we want to do is be able to grow as big as we can and invest in as much as we can and continue to increase the TAM. So we don’t have to think about pricing at this point, but we do know we have pricing power.
We actually believe and we try to say this on the last couple of calls which is the more we invest in the product, the more content, the more formats that we deliver that content, the more subjects, the greater the Q&A that we add. That the TAM is probably two to three times the size just in the U.S. alone about original number we gave out. And that’s because, it will cover the other 50% of subjects that we don’t cover. And it will go deeper in the subjects that we do cover. So we want wide at first now we’re going deeper. So we’re freshmen all the way through the seniors on key majors and that was not a year ago, the plan.
…our ability to integrate Wdesk with more than 100 cloud, SaaS, and on-premise applications including Oracle ERP Cloud has expanded our TAM to $16.2 billion.
We continue to gain market share in the SEC compliance market, where Wdesk is considered a best practice. We’re also seeing more demand for Wdesk from foreign private issues that use IFRS taxonomy because they’ll be required to submit their SEC filings with XBRL tagging starting December 15.
We finished Q3 with 2,991 customers, a net increase of 295 customers from Q3 of 2016, and a net increase of 83 customers from Q2 2017. Our subscription support revenue retention rate excluding add-ons was 96.5% for the month of September 2017, compared with 96.1% in June 2017, and 95% in September 2016. Customers being acquired or ceasing to file SEC reports, accounted for a majority of the revenue attrition, consistent with our experience to date. With add-ons, our subscription support revenue retention rate was 108.6% for the month of September 2017, compared with 106% in June 2017, and 108.7% in September 2016. Increased subscription revenue on non-SEC use cases from existing customers continues to be the primary driver of our add-on revenue retention rate.