Snap Inc (SNAP) 2017 Q3 法說會逐字稿

Good afternoon, everyone, and welcome to Snap Inc.'s Third Quarter 2017 Earnings Call. (Operator Instructions) This call will be recorded. Thank you very much.

Mr. Arman Panjwani, Investor Relations, you may begin.

Thank you, and good afternoon, everyone. Welcome to Snap Inc.'s Third Quarter 2017 Earnings Conference Call. With us today are Evan Spiegel, CEO; Imran Khan, Chief Strategy Officer; and Drew Vollero, CFO.

Earlier today, we made a slide presentation available reviewing our key engagement and financial metrics for the third quarter of 2017, which can be found on our Investor Relations website.

Now I will quickly cover the safe harbor. Today's call is to provide you with information regarding our third quarter 2017 performance in addition to our financial outlook. This conference call includes forward-looking statements. Any statement that refers to expectations, projections or other characterizations of future events, including financial projections or future market conditions, is a forward-looking statement. Actual results may differ materially from those expressed in these forward-looking statements, and we make no obligation to update our disclosures. For more information about factors that may cause actual results to differ materially from forward-looking statements, please refer to the press release we issued today as well as risks described in our prospectus dated March 1, 2017, particularly in the section titled Risk Factors. This information can also be found in our other filings with the SEC when available.

Our commentary today will also include non-GAAP financial measures. We believe that the use of these non-GAAP financial measures provides an additional tool for investors to use in evaluating ongoing operating results and trends. These measures should not be considered in isolation from or as a substitute for financial information prepared in accordance with GAAP. Reconciliations between GAAP and non-GAAP metrics for our reported results can be found in our press release issued today, a copy of which can be found on our website at investor.snap.com.

At times in our prepared comments or in response to questions, we may offer additional metrics to provide greater insight to our business or our quarterly results. This additional detail may be onetime in nature, and we may or may not provide an update in the future on these metrics.

With that, I'd like to turn the call over to Evan.

Good afternoon, and thank you for joining our call. As we are rapidly approaching the end of 2017, I thought it might be useful to speak to our progress against the 3 priorities I shared earlier this year: performance, quality and automation; and discuss our 3 new priorities for 2018: user growth, content and augmented reality. We have been very focused on making progress against performance, quality and automation this year and are beginning to see the results of these efforts. Application performance has increased considerably, and we have made meaningful progress against key customer-facing metrics. For example, since April, we have reduced the average time it takes for Android users to start our camera application by over 20%. Camera start-up time is very important for Snapchat because people use our service to quickly capture moments to share with friends.

As part of our quality efforts, we have been building a new world-class device lab for test automation and quality assurance. We've also distributed a wide variety of Android handsets across our engineering organization to ensure that we have ongoing qualitative feedback on Android application performance and design. In Q3 alone, Snapchat was used on over 60,000 different Android model variants. Given the sheer volume of different Android handsets used to access Snapchat, we have had to establish new processes to ensure that our quality efforts can be maintained. This will be an ongoing investment, but I am pleased to say that this quarter, monthly crash rates across both iOS and Android reached an all-time low since we started tracking this metric early last year.

Our efforts at automation have gained traction very quickly this year, with 80% of Snap Ad impressions delivered programmatically in Q3, up from 0% 1 year ago. The speed of this transition surpassed our expectations, but has dramatically reduced pricing as advertisers moved from direct sales to our unreserved auction. This has decreased CPMs more than 60% year-over-year, which has made it harder to grow revenues at the rate we would have liked. I am grateful that this transition is nearly behind us and look forward to the many advantages our programmatic auction brings to our advertising business in terms of scale and ROI. The number of advertisers spending in our auction grew nearly 5x from the beginning of the quarter, one of the many early indicators that our self-serve tools are making it easier for more advertisers to reach our audience. I'm excited about the case studies that Imran will share with you a bit later in the call.

The substantial decline in average CPMs during our transition to the auction has meant that the majority of our revenue growth has come from a dramatic increase in impressions. This has been a good test for our business because we are always concerned that an increase in ad load will negatively impact user engagement. Fortunately, we have seen user engagement continue to grow at a meaningful clip with time spent, frequency of use and Snap creation all increasing, while Snap Ad impressions have grown over 400% year-over-year. This increase in ad impressions, combined with the simultaneous growth in user engagement, bodes well for the long-term success of our advertising business. Of course, overall ad load remains very low, and we will continue to monitor user engagement as ad load increases over time.

I am proud of our team for executing relentlessly against our priorities in 2017, and I feel that we have built a solid foundation for the future. Looking ahead to 2018, our team is focused on user growth, content and augmented reality.

This quarter, we grew our daily active users at a lower rate than we would have liked, adding 4.5 million new users. This can be partially attributed to our decision to report our daily active users as an average over the entire quarter, where a strong September was offset by the more modest months of July and August. Ultimately though, we want to drive more user growth in 2018.

We are fortunate to grow our business from a position of strength with a young audience in a developed world, the tip of the spear that drives broader adoption trends across the technology ecosystem. For example, we now reach over 70% of the 13- to 34-year-old population in the U.S., France, the U.K. and Australia. However, in order to further scale our user base, we need to accelerate the adoption of our product among Android users, users above the age of 34 and users in the rest of world markets. This means that we will have to make some changes to our product and business.

To attract more Android users, we are building a new version of our Android application from the ground up that we will launch in select markets before rolling it out widely. This new version of our application leverages everything we have learned about building for Android over the past 5 years to provide a more performant product experience that we know our community will appreciate. This effort requires significant engineering resources across all of our engineering teams and will be a huge focus over the coming year. After seeing the results of increasing Android performance over the past few months, with significantly more Android users added than iOS users in September, we wish we had done this sooner.

The one thing that we have heard over the years is that Snapchat is difficult to understand or hard to use, and our team has been working on responding to this feedback. As a result, we are currently redesigning our application to make it easier to use. There is a strong likelihood that the redesign of our application will be disruptive to our business in the short term, and we don't yet know how the behavior of our community will change when they begin to use our updated application. We're willing to take that risk for what we believe are substantial long-term benefits to our business.

High-speed and affordable wireless connectivity is necessary for the use of our application, and this has proven to be an obstacle for growth in the rest of world markets. We are taking action, both internally and externally, to improve connectivity for our community. Internally, we are focused on product improvements like our new streaming architecture for story playback. This updated architecture means that we no longer have to download an entire video file before we initiate playback, which ultimately results in a better user experience and more efficient use of the network.

Externally, we are exploring partnerships with select wireless carriers who can help us provide our service at a lower cost to our community.

Content is becoming an increasingly important part of our business as many of our early investments and partnerships have begun to bear fruit. It has been nearly 3 years since we launched Publisher Stories, and I have been so excited to see the tremendous progress we have made together with our partners. Additionally, Our Story has grown from a one-off event-based product to an always-on content experience, discoverable first in Stories and now in Search and Maps. Our coverage of the recent hurricanes in the United States demonstrated the breadth of the Snapchat content experience, with unique perspectives across Publisher Stories, Shows, Our Story and the Snap Map. Over 3,500 hours of storm-related snaps were submitted to Our Story, providing an intimate and near-realtime perspective of those devastating events. We heard that many users that had evacuated the area were able to keep an eye on their neighborhoods using Snap Maps.

Professional journalism continues to play an important role on Snapchat and provides an important context to user-submitted snaps. Stay Tuned, a twice-daily show produced by NBC, reached over 12 million 13- to 24-year-old viewers in the United States in September, making it one of the most popular new shows for young people in the country almost overnight.

As part of the redesign I mentioned earlier, we're going to make it easier to discover the vast quantity of content on our platform that goes undiscovered or unseen every day. We think that there is a big opportunity to service some of this content in a personalized and more relevant way while still maintaining the exploratory nature of our service. We are developing a new solution that provides each of our 178 million daily active users with their own Stories experience, leveraging the tremendous benefits of machine learning without compromising the editorial integrity of the Stories platform that we have worked so hard to build.

As part of our efforts around Search and Maps, we now index millions of stories every day, meaning we have the long tail of content necessary to provide a truly personal experience. We hope that showing the right stories to the right audience will help grow engagement and monetization for our partners and for Snapchat.

While we have made significant progress in our work to empower the creation of user-generated content from friends and premium content from publishers, we have historically neglected the creator community on Snapchat that creates and distributes public stories for the broader Snapchat audience. In 2018, we are going to build more distribution and monetization opportunities for these creators in an effort to empower our creative community to express themselves to a larger audience and build a business with their creativity. Developing this ecosystem will allow artists to transition more easily from communicating with friends, to creating stories for a broader audience, monetizing their stories and potentially using our professional tools to create premium content.

Snapchat has likely become the world's most used camera, with more than 3.5 billion snaps created every day, up more than 40% year-over-year. According to InfoTrends, this is now greater than the average number of photographs taken every day this year across all phones, tablets and digital cameras combined, excluding Snapchat. The strength of our close friend network that drives high-frequency engagement means that our community opens our application an average of 25 times per day, creating numerous opportunities to inspire creativity.

Augmented reality is one of the many ways that we inspire our community to create snaps. With a tap of a finger, the Lenses transform the world around you and make even the most boring of situations infinitely more snappable. In September, we released 3D Bitmojis, which have increased the use of World Lenses by more than 20%. In October, we collaborated with Jeff Koons to place installations of his iconic sculptures in locations around the world. And more recently, we've helped the NBA create unique characters to bring team spirit to their arenas around the United States.

With all of the excitement and creativity inspired by each new lens on Snapchat, we are working hard to democratize lens creation so that anyone, anywhere can create and publish their own Lenses. We have been testing our Lens Studio product with a select group of advertising and creative partners, who have already created amazing Lenses with our tools. We'll be investing in improving our platform and tools as well as making our Lens Studio more widely available to empower the creativity of our community, much in the way that we have with our Geofilters product. With over 5 million Geofilters submitted in the past 2 years, we can't wait to see all of the Lenses created by our community.

Context Cards represent the next generation of our augmented reality products, with the information and actions overlaid on the content that our community is watching. We will be building on top of the Context Cards framework with additional partners as we learn more from our community about how they want to use this new product. With all the snaps and stories viewed on our service every day, Context Cards provide a unique opportunity to translate what you see on the screen into action, whether online or in the world around you.

2018 promises to be a productive and exciting year for Snap, with many changes coming to our products and platform. We will be hard at work delivering on our priorities: user growth, content and augmented reality.

With that, I'll turn the call over to Imran to discuss our business.

Thank you, Evan. Total advertising revenue for the quarter was $204 million, an increase of 59% year-over-year and 16% quarter-over-quarter. This represents healthy growth, but we are working tirelessly because we want to grow revenue faster. I want to focus my time today on a few factors that we believe have set us up well for strong growth in the future quarters: one, transitioning the majority of our business to self-serve; two, addressing the needs of our increasingly diverse advertiser mix; and three, democratizing our sponsored creative tools so that all advertisers can use them. I'll talk about how each of these factors challenged us and what we are doing to address them.

First, we have been focused over the past year on transitioning the majority of our Snap Ads business to self-serve. We launched our Ads API a year ago. This marked the beginning of our transition to programmatic advertising. We followed it in June with our own self-service tool. The transition to self-serve is an important part of scaling our business for a number of strategic reasons. For instance, in our auction platform, advertisers can now specify and optimize delivery against certain business objectives, such as driving app installs or video views. With this framework, we can address the needs of many more advertisers and help them achieve their desired business goals at scale. We have also made it easier for advertisers to test and learn. The auction has lowered the entry price point by 3 orders of magnitude, making it more accessible to all advertisers. And our self-serve tools have made it possible to manage campaigns with ongoing testing and realtime improvements.

Given these strategic reasons, we have been very aggressively transitioning our Snap Ad or full-screen mobile video business to self-serve, and advertisers' adoption has been tremendous. Today, over 80% of Snap Ad impressions are being delivered programmatically. This is up nearly 3x from Q1 of this year and up from 60% since last quarter. However, our auction has a lower price point than our reserved business because there is no fixed rate card. As we transition more and more of our business to the auction, this had a meaningful impact on overall pricing. While this diminishes revenue in the short term, it builds the foundation for long-term scalable revenue. As we onboard more advertisers and multiple advertisers compete for the same ad impression, we should see higher pricing. In fact, we are already seeing this happen in some auction segments. The number of advertisers spending in the auction is up nearly 5x since the beginning of the quarter. This has resulted in sequential growth in the number of contested auctions, and our data shows that these prices are, on average, higher than uncontested auctions. Increasing the number of advertisers on Snapchat is foundational to building an efficient marketplace, and our self-serve tools help us do this at scale. More advertisers means a greater variety of ads, which is a win-win situation for both our advertisers and our community because it helps us deliver the right ad to the right person. We're confident that increasing the number of advertisers, combined with our ongoing machine learning efforts, will lead to better results for our advertisers, better experiences for our community and higher pricing for us in the long term.

I would like to share an example of how our self-service auction platform is helping our customers achieve their objectives. Peak Labs, a U.K.-based brain-training app leveraged our auction for an app install campaign. They developed over 10 different Snap Ad creatives to keep the ads fresh and engaging. Using the auction's realtime reporting capabilities, Peak was able to rotate creatives on a biweekly basis, analyze performance via A/B tests and optimize efficiency for the overall campaign. Results were great. After 2 months, Peak drove 1.7 million app installs at a cost per install that was 50% lower than other platforms where they were advertising.

The second area of focus for us is addressing the needs of our increasingly diverse advertiser mix. As we onboard more advertisers, it is critical that we understand the nuances of each vertical and customer type. Last year, we were focused on large advertisers. Over half of our revenue in 2016 was tied to large share of voice or sponsorship deals. Many of these customers wanted third-party measurement solutions, and as a result, we invested heavily in this area. We have made good progress. We now have over 18 third-party measurement partners, and more than half of every dollar spent on our platform has third-party measurement attached to it.

Here's an example of how our third-party measurement helped an advertiser achieve their business goals. 2K was looking to drive awareness and purchase intent of their NBA 2K18 video game using Snap Ads and Lenses. Through our partnership with Nielsen, they were able to measure the results that mattered most. Among users who were exposed to both the Snap Ads and Lenses, they achieved a 45-point lift in awareness and a 24-point lift in intent to purchase the game.

Additionally, our acquisition of Placed represents our commitment to open-source attribution. Over 90% of retail transactions still happen offline in the physical world. Placed standardizes the omnichannel measurement of store visits and enables advertisers to understand performance that goes beyond a walled garden. It ensures that advertisers can understand performance in the context of their entire budget. In Q3 alone, Placed measured campaigns for more than 300 advertisers. The vast majority of businesses in the United States are small- to medium-sized and don't have access to television advertising or Madison Avenue. With the launch of our self-service platform, we are investing heavily in expanding our existing advertiser base to include long-tail and mid-tail advertisers.

We're seeing great traction. We more than tripled revenue from SMBs in the third quarter when compared to the first half of 2017. But to unlock the true potential of this segment, we need to invest in first-party measurement solutions. As such, we've recently begun to roll out additional attribution capabilities. In Q3, we began a small alpha of Snap Pixel, our conversion SDK, with a handful of advertisers. Snap Pixel allows advertisers of all sizes to track the impact that Snap Ads have on online sales, lead generation or other conversion actions in realtime. This is particularly important for mid-market and SMB advertisers. Additionally, the Pixel will soon support the targeting of custom audiences based on on-site behavior and bidding based on conversions, not just impressions or swipes.

One of our first advertisers to test Snap Pixel was TechStyle, the parent company to Fabletics, JustFab and ShoeDazzle. They focused on 2 KPIs: cost per sign-up and cost per purchase. It's early, but the results are promising. Across all their brands, they saw a cost per purchase that was about 40% to 60% lower than their goals within 24 hours after someone saw their ad. The retailers also achieved a cost per sign-up that was roughly 30% to 50% lower than their goal. They were able to attribute 28% more conversions from their Snapchat ads by looking at 1 days’ worth of view-through attribution.

The third area of focus is democratizing our sponsored creative tools. Over 60% of our daily active users create snaps with our camera every day. Sponsored creative tools, which includes Lenses and filters, allow advertisers to leverage this unique creativity to reach their audiences. Until recently, our Sponsored Lens product has been accessible primarily to larger advertisers looking to maximize reach. Based on comScore Xmedia, the average audience of the top primetime shows is between 5 million to 12 million viewers. A National Lens on Snapchat can reach well over 20 million users in 1 day. Because the user experience remains our top priority, we limited Sponsored Lenses at one per day. This was the right decision as we refined the ad product, but it also impacted our ability to scale. We addressed this in the second quarter with the launch of Audience Lenses. Audience Lenses enable advertisers to purchase Sponsored Lenses that reach specific audiences at a better entry price point. Advertisers can use the same targeting capability that they already use for Snap Ads, including age, gender, geography and custom audiences. We're confident that this will help reduce seasonality and globalize our creative tools business. In fact, the number of lenses sold in Q3 was up nearly 15% from the last quarter and more than doubled the amount sold in Q3 of last year.

Let's take a look at an example. Adidas wanted to target high school and college athletes for their Here To Create campaign. They used custom audiences to reach this particular demo, and now with Audience Lenses, we were able to do this across all of our products. This allowed them to amplify their message to the most relevant audience, and results were fantastic. They saw an 18% increase in new visitation to their retail stores. We recognize the areas we need to invest in, and we're acting quickly. Much of the work that we have done this year is laying the groundwork for 2018 and beyond, but we're encouraged by what we have seen so far.

Thank you, everyone, for the time, and now I will pass it to Drew to discuss our financial highlights for the quarter.

Thanks, Imran, and good afternoon, everyone. Snap continues to make solid progress against its long-term goals, and the growing traction across many areas of our business is encouraging. Let me speak to the third quarter highlights. First, both Snap Ads and creative tools remain a strong one-two combination. Revenue from both product lines increased sequentially and year-over-year. We also saw more growth in engagement in the quarter. Metrics such as time on the app, video views and frequency of use all grew in Q3.

We made meaningful progress building out our auction platform. Numerically, over 80% of Snap Ad impressions were delivered via the auction, up from 60% in Q2. Overall, Snap Ad impressions grew substantially. Snap Ad impressions increased over 60% sequentially and over 400% year-over-year. This means that we were able to grow ad impressions and engagement simultaneously, which is a positive sign for the long-term growth of our business.

As Imran mentioned, the auction transition continued to impact Snap Ad pricing in the quarter. In Q3, Snap Ad pricing was down more than 20% sequentially and over 60% year-over-year, most of which was driven by the mix shift from reserved inventory at rate card prices to the unreserved auction. On a positive note, we were able to partially offset pricing declines with auction contestation. During Q3, we saw that auctions with multiple bidders resulted in prices that were over 40% higher than uncontested auctions. Furthermore, prices for ads that were contested grew sequentially in the quarter.

On the cost side, our gross margins continue to scale well. We're seeing clear leverage in hosting and revenue share expenses, which is driving continued gross margin expansion. Please note that when I discuss all of our expense figures, including gross margins, they will exclude stock-based compensation and related payroll taxes as well as depreciation and amortization and nonrecurring charges.

Gross margins grew to 21%, up 230 basis points sequentially and over 2,000 basis points versus a year ago. Hosting cost per user have risen from $0.64 a year ago to $0.68 in the quarter, only a 7% increase over the last 12 months. We have seen revenues and engagement grow at much faster rates, creating cost leverage. We have excellent traction on some of our hosting cost initiatives and are benefiting from the dual-cloud hosting environment.

Similarly, as expected, we continue to see leverage on partner rev share expenses, driven primarily by monetizing My Story. Revenue share expenses as a percentage of sales declined to 10%, down 300 basis points sequentially and 500 basis points year-over-year.

Geographically, North America remains a leading region for ARPU and gross margins. In the third quarter, North American ARPU was $2.17, which is 7x higher than the rest of the world, but is still well below our peer set. Periodically, we analyze our hosting cost by region, which considers factors such as regional differences and app usage and varied regional hosting rates. This internal analysis estimated that our North American gross margin was already over 50% in Q3.

Switching gears. Operating expenses were thoughtfully managed. In the quarter, costs decreased 2% sequentially to $223.1 million. We continue to manage back-of-house expenses closely, and corporate G&A fell $8 million, driven by lower employee expenses and reduced legal costs. We also benefited from the timing of certain expenses, including lower trade show costs. The annualized cost per employee was slightly over $300,000, which compares favorably to our peer set. This is down versus prior periods.

Let me provide a brief update on a few of the investments Snap has made. One of Snap's biggest investments has been in our team. We have overinvested early to achieve the scale necessary to compete globally. At the end of Q3, we had nearly 3,000 employees, 80% of which are in the front of house. In the quarter, we added almost 250 people organically, consistent with our expectations. Additionally, we added nearly 100 people with the Placed acquisition.

Capital investments in the quarter were modest again at $26 million. Total CapEx per user was what we believe to be an industry-leading $0.15. For M&A investments, we completed the Placed acquisition in the quarter, which is off to a good start. Total consideration was $139.6 million.

Unfortunately, we misjudged strong early demand for Spectacles and purchased more inventory than we now anticipate being able to sell. As a result, we recorded a $39.9 million nonrecurring expense, primarily related to excess inventory and purchase commitment cancellations. Moving forward, we will continue to be in the marketplace with Spectacles and expect modest revenue from the product line.

We remain in a strong cash position and ended Q3 with $2.3 billion in cash and marketable securities. Capital deployment priorities remain business operations first, followed by opportunistic M&A. Since the IPO, we've been net-settling employee shares for tax purposes. In Q3, we used $162 million to satisfy tax withholding obligations. Moving forward, we're planning a transition to sell shares on the open market on behalf of employees for tax purposes. As of September 30, 2017, total shares outstanding were 1,202,000,000 and 1,441,000,000 on a fully diluted basis.

As we move forward, we want to share thoughts on the short term. We will continue to scale the auction business. We plan to continue the transition of reserved impressions to the auction, and we exited the third quarter with higher than the Q3 average of 80% flowing through our auction platform. As a result, we're planning to see similar auction dynamics to Q3, specifically, gains in impressions and lower prices. We look forward to completing the transition of our Snap Ad business to the self-service and auction platform.

On hosting expenses, we ended the quarter with slightly higher costs per user than the quarter average of $0.68. Historically, we have seen hosting costs increase during the fourth quarter due to the higher usage over the holiday season. We have seen a similar trend in operating expenses. Historically, our operating costs per head have increased low double digits from Q3 to Q4. As the team continues to settle in, we expect gains in employee productivity to be meaningful. This should allow us to reduce the scope and pace of hiring. Priorities for hiring will continue to be in engineering and international functions. Given this, we expect the pace of hiring to slow in the short term, particularly in the seasonally slow fourth quarter.

With respect to stock-based compensation and related payroll tax expense, our Q3 expense was $225.6 million. In the short term, we believe this is a good proxy for future quarters. Future acquisitions will likely be additive to this amount.

With that, I will now turn the line back over to the operator, who will open up the call for questions.

(Operator Instructions) Our first question comes from Stephen Ju with Crédit Suisse.

So Imran, in order for the auctions to continue to be highly contested, it seems like you have to give advertisers ongoing reasons to pay more because that is presumably more highly targeted and has higher ROI. So it sounds like you're just starting to ramp up the attribution tools. So any way to characterize what percentage of your typical advertisers' budgets' are being spent on a more highly targeted basis?

Yes. I think a couple of things need to happen for auction density to grow. So first off is onboarding a lot more advertisers. And I think, remember, we launched our self-service platform on -- in Q2, and we are pretty excited of the progress we made. We talked about it. We saw ibex increase the number of advertisers on our platform, and we continue to invest heavily on the inside sales team, reaching out to small marketers and onboarding them to more and more advertisers on the platform. At the same time, in terms of targeting, we actually made pretty significant investment over the last 12 months to help people reach the right -- show the right app to the right audience. Because when you show the right app to the right audience, it actually drives significantly more value for our users, for more of our advertisers. And I think we made pretty good progress on that. On the -- first, for the attribution side, Pixel is out. I think we're going to continue to push it forward. And I think that will drive advertisers to even more visibility in terms of what kind of ROI they are generating.

Our next question comes from Mark Mahaney with RBC.

This is Dylan Haber on for Mark. In terms of your programmatic advertising business, what are your views on opening up the Snap platform to third-party DSPs and exchanges to add incremental demand?

Yes. I think right now, we think the best way to drive our business is to have the self-service platform where advertisers can reach directly rather than the DSP. We are -- we have an incredible audience base, and we are very, very excited of that audience and understanding about the audience. I think when you open up the DSP platform and things like that, there's always a risk, that information leakage out of your platform, and that's not optimal. So I think we like our strategy of having a self-service platform where our advertisers and agencies can come and reach the audience they want to reach and delivering the most right app to the right person.

Our next question comes from Lloyd Walmsley with Deutsche Bank.

One for Imran and Drew, and I guess, one for Evan, if I can. First, you guys talked about the pricing declines of the shift to programmatic as a headwind to revenue, which seems obvious on a pricing impact. But are you really seeing budget pressure from this? And specifically, can you talk about what clients generally do with budgets when they shift to programmatic? Are they buying a consistent volume of impressions and shrinking budgets, or just reinvesting to get more volume such that it isn't really that much of a headwind to revenue? Any color you can share there would be great. And then a strategic one for Evan. It seems like a lot of changes focusing on user growth, focusing on the creator communities. So just wondering if you can elaborate a bit on how you guys have come around and how you get comfortable really that this won't disrupt the best friend engagement that you guys have historically been so focused on?

I think in terms of advertiser adoptions, you have seen that roughly 400% growth in ad impressions on our platform on a year-over-year basis. So we are very pleased with the number of ad impression growing on our platform. And I think as Evan pointed out in his script, despite the significant impression growth, we saw engagement on the platform grew on a year-over-year basis. I think one of the key things also self-service is allowing us to do is bring in a lot more advertisers on and use our advertising platform. We really didn't get the dollar from the direct response advertiser or performance marketers and a lot of mid-sized advertisers. And by opening up the self-service platform, we are bringing them on the platform. I wanted to share an example of a small advertiser who is not necessarily a performance advertiser, would not be successful on our platform if it was just reserved buy. So GOAT, which is an online marketplace for buying, selling and sneakers -- buying and selling sneakers on mobile, they used our platform to drive app install. And they were able to drive LTV, lifetime value, 20% above their goal and improve their payback period by a month compared to other marketing channels. So this is a perfect example that by opening up the platform, we were able to attract advertisers who would not be, in the past, successful on our platform and they can spend more money and drive their goals.

And you're right to point out that this communication between close friends is really valuable to our business. And it remains the most important thing for Snap. So I think the really exciting thing about the redesign coming is that we found a way to preserve and I think, in many ways, enhance that friend communication while still providing more opportunities to provide distribution and monetization opportunities to contact freighters. And so I think we're excited to roll that out, but you're right that our focus here really is preserving that frequency and intimacy of communication between close friends.

Our next question is from Ross Sandler with Barclays.

2 questions. Evan, first, on the user base, you called out that September was a big month for Android, it added more net adds than iOS. So do you expect that to potentially accelerate going forward? And you talked about how you're working to simplify the app. Is that happening right now? Or is that something that could happen in the future? And then Imran, on the self-serve platform, can you just talk about the behavior that some of these new advertisers who are signing up in self-serve for the first time, what is the behavior? Are they ramping up their spend each week? Are they coming in and spending and then dropping off? Just any general color on the behavior of self-serve, that would be great.

Yes, so given the distribution of Android devices in the world, I would expect that over time, we will see more net adds coming from Android, which is why we focus so much on improving the Android experience and why we're rebuilding the Android application from the ground up to make it easier to use. And right now, we're using a new version of the application internally and having a lot of fun with it. So excited to roll that out.

Yes. I think, Ross, with regards to your question, previously, before the self-service platform, if an advertiser wants to buy an advertisement, they have to commit a significant chunk of dollars, in many cases, a couple of hundred thousand dollars. And that's really difficult for many, many advertisers to come invest on a right -- on a new platform. With the self-service, we saw that many advertisers are coming in on our platform and they're testing. And as they are seeing success, they're increasing their budget. I think one of the key things we have seen, that our performance advertising dollar or DR dollar has grown significantly. Also, I talked about in the call that revenue from our small- and medium-size businesses has tripled. So we're seeing great traction with that as advertisers are coming and testing and learning and they're spending more.

Our next question is from Brian White with Drexel.

Evan, you mentioned the ARs are priority in 2018. And I'm wondering, you put a pretty cool World Lens out there, Stranger Things with Netflix, and I'm wondering how that innovation was received. And also, how are you thinking about leveraging the TrueDepth camera system on the iPhone X. I know Craig Federighi did a demonstration at the Apple event, September 12, and did a shout out for Snap. So if you could just expand on that.

Yes. Thanks, Brian. I wish I had more data for you around the Stranger Things lens because I agree that was super cool, and we're excited to see more and more advertisers experimenting with those products. I think they're creating really unique experiences for our community. So that's a win-win for us when our community really loves something and advertisers are creating and spending a lot of time interacting with it. And as for the Apple innovation, it's very exciting to see, I think early days for Apple and their investments around AR. But the TrueDepth camera obviously is great for Snap and for the -- especially in the front-facing camera. Those Lenses, I think, are enhanced by that building block that they provide.

Our next question is from John Blackledge with Cowen.

2 questions. On the app redesign, if you can give us a sense of the timing, any more insight into the changes and why you think the redesign could be disruptive in the near term? And then on the user side, just given the redesign and the new Android app, how should we think about user growth in the near term?

So I don't have any timing to share with you today, so -- and we'll, I guess, look forward to surprising you with the redesign. But I think conceptually, we've been spending a lot of our time sort of studying the evolution of content feeds on mobile. And so if we kind of go back to the beginning of content feeds, Twitter was really the first content feed on mobile and it was intra-space. So you would follow things you're interested in or news organizations or celebrities. And obviously, the next evolution in that content feed, with Facebook's invention of a content feed based on content curated by your friends. And this was a very interesting innovation because it obviously personalized the content feed based on what your friends thought you would be interested in. And I think there's a really exciting opportunity here for another evolution of that content feed that addresses some of the shortcomings of the friend-based content feed model. So for example, on a friend-based content feed, in order to get more content in that feed, you need more friends. And when people start adding more friends, they then feel less comfortable posting content, and so they start posting less. And that means that you need even more friends to get more content. And so you end up in this kind of precarious situation where because you based the content feed on what friends are posting, you are sort of inherently limited in how you grow that selection of content. And ultimately, what we found is that the best predictor of what people are interested in and want to watch is actually what they're watching, right? And I think there's an opportunity here for us to create a really great personalized content service that doesn't at all diminish the great and I think, very differentiated communications business that we've established.

Our next question is from Justin Post with Bank of America Merrill Lynch.

My question is really the mix of time on the site. Clearly, snaps are up 40%. Any change in that mix? And then can you talk a little bit about the monetization of snaps, how that compares versus maybe the Discover tab or Stories. How do you think about that?

Sure. So we don't break out that mix of time on the service. Overall time spent has been growing, which is something that we're really excited about. And so hopefully, we can share more with you in the future around that. I think it is a great opportunity to talk a little bit about our creative tools like we talked about earlier around Lenses because those are a really fantastic way that we monetize our communications service. And so not only are we able to monetize the consumption side of the business around Stories, but also with creative tools like Geofilters and Lenses. That has proven to be a very effective way, as Imran shared, to drive monetization around our communications products.

Yes. I think one of the key things to point out on Lenses and filters, I talked about in the call how we saw that because of Audience Lenses, we saw 15% growth in the number of Lenses sold on our platform in Q3 and doubled year-over-year. The other interesting thing is these are a very interesting product and a completely differentiated product. For the longest period of time, advertisers are trying to figure out how can we be part of the consumer conversation without being intrusive. Our Lenses product and filters product actually give the advertisers the ability to be part of the consumer conversation and in a very premium fashion. And the other interesting thing is that the amount of time people are playing with those Lenses product, actually pretty significant. In a world in mobile where the attention span is significantly lower, the playtime on Lenses are actually very, very exciting to many, many advertisers, and that's also driving the adoption. So I think we're actually very, very excited about both of our products, Snap Ad and our creative tools, and we think there's opportunity to monetize those product long-term.

Our next question is from Brian Nowak with Morgan Stanley.

One just on advertising, kind of a high level picture on the ad format. Ad performance really matters to driving ad budgets. You're talking about a redesign of the app. How do you think about the potential risk that a minute spent in Stories format is just going to monetize materially lower than other social interaction like a news feed. Is that a risk just because of the way people consume Stories, you may not be able to monetize that as high as other formats? And if that's the case, should you be adding a news feed?

I don't think we would add a news feed per se, but I definitely think there are ways to improve ad performance on our service, and we're always looking at evaluating that. And as part of the redesign, I do think there will be new monetization opportunities. Again, too early to tell there, but we're excited about it.

Our next question is from Douglas Anmuth with JPMorgan.

I just wanted to hit on 2 things. First, you talked about the DAU trajectory through 3Q, and then it kind of ended stronger in September. Just curious along those lines what you can tell us about Maps and how that's doing. It didn't seem to get as much commentary tonight as it did 3 months ago. And then secondly, just how that DAU trajectory is kind of early in 4Q. And then just on the advertiser side, could you just help us understand when an advertiser would use the direct sales force going forward versus the auction format?

Yes. So on Maps, very excited about how that product is progressing. And we continue to really invest in this action-emoji concept, which is very cool because it allows people to express themselves without creating a snap. So as we look to empower self-expression and remove the friction from self-expression, sometimes, people are in the moment and they don't necessarily want to create a snap, and the Map allows them to show that, for example, they're on a bike and listening to music and their friends can interact around that behavior. So as far as the Map is concerned, very excited about it, but still hidden behind a pinch gesture. And we look forward to elevating the Map in our products so that more people can find it easily. Obviously, I'm not going to provide guidance on DAU, but I think it is important to note, our audience may not be the largest today, but we certainly feel that it's the most strategic and we're very excited about the 70% of 13- to 34-year-olds that we reach in the U.S., U.K., France and Australia. And I think that, that is a really strong base to grow from because it's very easy to monetize, and we'll be able to fund our future growth in countries that may be harder to monetize in the short term but that we believe we'll monetize in the longer term. So as I mentioned today, we are taking steps to change our product and change our approach to grow outside of those key markets where we do think that there are future monetization opportunities. But we really wanted to make sure that given the strategic nature of the 13-to-34 audience that, that was an audience that was really engaged with the snap platform before we expand it beyond our core.

Yes. I think with regards to direct sales force versus auction, I think it's really important to understand that auction is a buying mechanism. It lets our advertisers to come in and buy, test and learn, redirect their campaign and see what drives better results. The value of our sales force is incredibly high, primarily with the auction because the advertisers need to understand our platform, having insight on what works on our platform, how to win on Snapchat, those kind of value that our sales force provide to our clients. So as what we were saying, that as we make our buying platform more and more automated, our sales force transition from taking orders to become more of a consultant for our clients to become the partners. So that frees up their time and they can focus more and more time solving the client's business problem because when you solve the client's business problem, that drives really success. So I'll share one quick example. So for example, Activision, they wanted to test a short form of ad on our platform. So they created these 5-second snaps on our platform. They worked with our sales team and for their Call of Duty: WWII title. And that was a 5-second cinemagraph ad. They worked with our creative strategy team. And that drove 19 points lift in brand awareness because that our sales force had more time to work with the clients and give more insight, and that really drove us to this win-win situation. So I think that's one of the -- another perfect example. So I think direct sales force will become an enabler and become a consultant for our clients.

Our next question is from Rich Greenfield with BTIG.

I guess just from a really high level, it seems like you've only been public for 8 or 9 months. It seems like a significant amount of change in a short period of time. I mean, the auction team sort of surprised you. You're changing the apps to make it easier for users. You're embracing influencers. I guess just when you look at the short period of time since the IPO, where you seemed pretty upbeat, what's happened so fast, like what led to such a significant shift? And I guess maybe attached to that is, I guess, for Evan, like in the senior team that you have assembled, not just on the call but overall, is your senior team, the team you need, do you need a much bigger team or a different team? And then just a housekeeping question, on Spectacles, I think at the Vanity Fair conference, Evan, you had commented that Spectacles were exceeding your expectations. Yet today, you're writing them off, like $40 million of inventory. I'm just wondering how you square exceeding expectations with the write-off would be helpful.

Thanks, Rich. You're right, there has been a ton of change in a very short period of time, and that's been the case over the life of the business. In the last 6 years we've been in business, we continue to evolve the product very rapidly. We continue to evolve the business very rapidly. And I guess we're just not afraid to make changes in the long-term interest of the business. So I would expect us to continue learning as we grow the business and making changes that we think are in the best interest of growing the user base, growing revenue and ultimately, providing our customers with a great product experience. As it relates to team dynamics, I'm very happy with our executive team, but I'm constantly evaluating our team, providing feedback, and we're all working together to grow and drive the business

(technical difficulty)

as productive as a way as possible. And when it comes to Spectacles, you're right, we were very excited about Spectacles and by the initial reception. And because we were so excited, we made, I guess, the wrong decision. And we were bouncing the trade-off with unit economics, of course, that come with hardware. But ultimately, we made the wrong decision based on the early traction and ordered a lot of long-lead time parts. And ultimately, we weren't able to sell as many Spectacles as we thought we'd able to base on the early adoption. So we're learning from it, and we plan on avoiding a similar mistake in the future.

Our next question is from Jason Helfstein with Oppenheimer.

2. First, just thinking about the brand advertisers, can you talk about how the annual budget cycle played into it? It seems more and more budgets are being based on a calendar basis versus, in the old days, it was more kind of on the school calendar. And kind of where you were at the beginning of this year, perhaps not in a position with measurement, some of the things that advertisers are demanding, and kind of how that perhaps handicapped you this year, but then how you stand going into next year as those budget decisions are being made? And then also, how do you derisk the redesign of the app, since we know that your younger users are quite sensitive to the app that was originally designed for them?

Yes. I think in terms of advertisers, upfront deals and things like that, look, I think one of the key things is to drive more upfront customer -- deals we have to drive value for our advertisers. And as our business is growing, we are able to show more and more value, we're able to show more and more products, more and more measurement, more and more targeting and also able to cover them for a while. A couple of years ago, I think one of the biggest complaints I heard from a lot of our agency customers that, hey, we don't hear from your team more often because we didn't have a big team. Now I think the investment we made on the headcount helped us to cover them and look at them and helped them to really show what success could look like on Snapchat and how to win on Snapchat. And I think that's driving and we are having more and more conversations about upfront and I think having a much more deeper conversation with the agencies on what we can do together. So I think we're making good progress. I'm pretty pleased with it. And on top of that, we're onboarding a lot more new advertisers we couldn't reach on our reserve platform because now it's opened up the opportunity to reach out to performance advertisers and small-, medium-sized businesses.

And as it pertains to the redesign, again, we don't know exactly how that's going to play out, but we will be testing the product, both internally, which is ongoing, and externally. And I think we're going to learn a lot from that. But fundamentally, as we look at the product evolution at Snap, we really try to build products around a thesis. And sometimes, that means that there isn't a strong early adoption of our product, like, for example, our iconic Stories products basically had no one using it for the first 6 months or so. And so I think, over time, we've learned not to be fearful of making big product changes that we think are in the best interest of our community, and we're going to keep going.

Our next question is from Mark May with Citi.

And I apologize if it's already been asked, but trying to get a little bit better handle on the impact that the growing auction-based ad platform is having and will have on financials that we all see every quarter. Kind of can you give us a general sense of the difference in the prices in that market today relative to your direct sold business? And also, what portion of the overall revenue is generated from it, and obviously, since you're trying to get to where are we in this transitional period as it relates to the trade-off between pricing pressure and then growing your advertiser base and...

Good. Thanks for the question. So at a high level, the auction is a strategic move for the company. It's something that's important for our business as we move forward, and we think it's the right way for us to be growing long term and to be broadening our reach. The dynamic that we saw in the third quarter with the auction are consistent with what we saw in the second quarter, and that is growth in revenue on the auctions platform, driven by significant increases in impressions. So we saw impressions up 400% year-over-year and up over 60% sequentially, so that impression growth is real. On the flip side, we did see as we moved from sort of a rate card pricing model to an auction-based model, we did see pressure in pricing. Sequentially, prices were down 20% in the quarter and 60% year-over-year. So that was real. We did see, where we see contestation and we see bidders, multiple bidders for that same advertising space, we did see prices that were 40% higher than places where we didn't have a contested auction. So that really bodes well for the future. Obviously, one of our key strategies here is to really grow the competition within the auction and then have more advertisers using the platform. That does bode well for where we are. So that's the auction platform. It's really the centerpiece of our Snap Ad business as we migrate there. We were -- 60% of our ads were sold on that platform in the second quarter. It's now over 80%. And as I talked about our commentary in the fourth quarter, it's going to be important as that transition continues. As we plan our business, we're thinking about that transition continuing, and we continue to see a bump in -- we're thinking we're going to continue to see a bump in impressions and continued pressure on pricing. So that's the Snap Ad business at a high level. We also have that good one-two punch with creative tools. The creative tool business is a good business for us, and we made progress there as well.

The next question is from Michael Nathanson with MoffettNathanson.

I have one for Drew and one for Evan. Following the answer to Mark's question, can you drill down a bit on your high-level view of the auction? Could you give me a sense of the rule of 80-20? So if 80% of the impressions were sold programmatically, what does that equal to in terms of advertising revenues as a percentage of revenues?

So at a high level, the auction is a growing piece of our business. We haven't broken out the differences between Snap Ads and creative tools. I think we've talked about a good solid one-two punch, so both of them are important pieces of revenue. So I think based on that, you can get to where you need to get to, is that the majority of Snap Ads' business right now is sold through the auction. More than 80% from an impression-based get sold there. So it's an important and growing piece of our business. But in terms of quantifying the actual amount, we're not going to do that.

And our last question will come from Youssef Squali with SunTrust.

2 questions. One for Imran and maybe one for Evan. Imran, remind us again that for the -- about the timing of the launch of the programmatic platform, was it Q4 of last year? Just trying to figure out exactly when do we lap that launch and when do we start having that maybe stop being a headwind? And then Evan, you launched Context Web -- I'm sorry, Context Cards last quarter. Can you speak to early learnings there and how this may help you leverage partners to supervise users with more information and especially, ability to transact? Maybe if you can address that, that will be great.

Yes. Thank you for the question. So with regards to timing, our API was the first partnership, was the first initiative to start the business programmatically. That was launched in October of last year. And then we expanded our API partnership in Q1. And then in late Q2, we launched our self-service platform. To give you a better sense as you're trying to understand the impact of this auction business over time, in my prepared remarks, I talked about it, that how auction volume for Snap Ad business as a percentage of overall Snap Ad business was over 80%, and that is up 3x from Q1 and up from 60% in Q2, just to give you some sense.

Yes. And as it pertains to Context Cards, I'm very excited about it. Initially, Context Cards are focused on places. And I think one of the coolest parts about Context Cards is they leverage this behavior that's sort of known across our ecosystem, where people swipe up on things that they're interested in to take an action, whether that's an ad or it's a content posted by their friends or content that they see in Maps. And so we're leveraging the sort of learned behavior of swiping up to get more. And obviously, today, it's only really around places, but we're very excited to expand that to learning more about all sorts of things, and that's something we're going to continue to invest in.

This concludes our question-and-answer session as well as the conference. We thank you for attending today's session, and you may now disconnect.