Good afternoon, welcome to the Twilio's Q3 2019 earnings conference call. My name is Daphne, and I will be your operator for today's call. At this time, all participants are in a listen-only mode. Later, we will conduct a question and answer session. I will now turn the call over to Andrew Zilli, Vice President of Investor Relations. Mr. Zilli, you may begin.
Thanks. Good afternoon, everyone, and thank you for joining us for our third quarter fiscal 2019 earnings conference call. Our results press release, SEC filings, and a replay of today's call can be found on our IR website at investors.twilio.com. Joining me today are Jeff Lawson, Co-founder and CEO, George Hu, COO, and Khozema Shipchandler, CFO. As a reminder, some of our commentary today will be in non-GAAP terms. Reconciliations between our GAAP and non-GAAP results and guidance can be found in our earnings press release. Additionally, some of our discussion and responses may contain forward-looking statements which are subject to risks, uncertainties, and assumptions. Should any of these materialize, or should our assumptions prove to be incorrect, actual company results could differ materially from these forward-looking statements.
A description of these risks, uncertainties, and assumptions and other factors that could affect our financial results are included in our SEC filings, including our most recent report on Form 10-Q, and our remarks during today's discussion should be considered to incorporate this information by reference. Forward-looking statements represent our beliefs and assumptions only as of the date such statements are made. We undertake no obligation to update any forward-looking statements made during this call to reflect events or circumstances after today, or to reflect new information or the occurrence of unanticipated events, except as required by law. With that, I'll hand it over to you, Jeff.
Thanks, Zilli, welcome everyone to this quarter's earnings call. Before I begin, I'd like to take a moment to recognize and thank the nearly 5,000 brave firefighters and other first responders doing their best to keep the fires in Sonoma County at bay, protecting the lives, homes, and livelihoods of our neighbors to the north in absolutely horrifying conditions. Thank you. Now, on to our call. Our third quarter results were strong, with total revenue growing 75%. Tremendous growth at this scale that reinforces our view that we are still in the early days of this market opportunity. While we are the leader in the category today, we have much bigger aspirations. We are investing across the business in new products, new regions, and much more to position us for long-term success to take advantage of this fast-growing market.
We're very confident that the demand environment for our customer engagement platform and the inputs to our business are strong. I recognize that base revenue came in slightly below our guidance, and Khozema will go into more detail on that. But that doesn't change our perspective that we've only scratched the surface and are excited about what lies ahead. We continue to help companies reimagine their customer engagement as every company needs to focus on building great digital experiences for their customers. Our commitment to innovation and customer success is driving relationships with companies of all sizes as the breadth of our product offering and our developer-first approach position Twilio as the platform of choice for building meaningful relationships with customers. I'm particularly excited about some of the new companies we've started working with in Q3, which George will talk about shortly.
Nowhere was this developer-first approach more evident than at our annual customer event, Signal, where we welcomed more than 3,200 customers, prospects, developers, and business leaders and introduced a number of new products and features. At Signal, I told you we had over 6 million registered developer accounts on our platform. Today, that number is now north of 7 million and climbing. Innovation has always been at the core of Twilio, and you can see that in the new products and features we announced at Signal, the biggest being Twilio Conversations. Everyone has had the experience of getting a message from a brand, only to find out that most times responding to that message doesn't work. We've reached a point where customers expect this kind of easy, frictionless engagement with brands.
At Signal, I mentioned some brands who have figured out the importance of meaningful two-way conversations with their customers, companies like Nordstrom, Macy's, Lyft, Redfin, and others. We've heard from many customers, especially as we launched Flex, that they see this as an opportunity to transform the way they engage with their customers. That's why we built Conversations, to make it easier for every company to build this kind of two-way engagement across channels and deliver a better customer experience. We also announced Ads for Twilio SendGrid, a new offering within our marketing campaigns product to address the more than $300 billion digital ad spending market. SendGrid Ads synchronizes across all of your email contacts and allows you to target ads across three of the major advertising platforms, Facebook, Instagram, and Google Ads.
This is all managed from a single UI, and more importantly, the ads are informed by the intelligence of the email channel so your ads can be smarter. We also announced the Email Validation API, which validates email addresses before companies send to them, decreasing bounce rates and improving sender reputation and inbox placement. There's always a tremendous amount of energy at Signal, and this year was no different. I want to thank our team and everyone who attended for making this the best Signal yet. Signal is a great example of the innovation that Twilio continues to deliver to our customers. Because of this innovation, focus on customer success, and our extensive customer engagement platform, Twilio was recently recognized as a worldwide leader in the IDC MarketScape for CPaaS, being named the de facto icon of the CPaaS segment for our vision and strategy to drive digital transformation.
The report highlighted how Twilio has almost single-handedly created a new communication segment, and specifically mentioned our recent focus on solutions such as Flex. This is great validation that our constant innovation and our vision of powering the future of communications is bearing fruit. Speaking of Flex, we're continuing to see great interest and traction with the product as we just passed the first year of GA. We're continuously innovating to add high-value features our customers are asking for. In fact, at Signal, we announced the Media Streams API to give companies access to the voice media of the calls coming into Flex, allowing real-time transcription, conversation analytics, and more. We also announced native CTI integrations between Flex and Zendesk and brought Autopilot to GA within the Flex console.
While it's still early for Twilio Flex, we are seeing some great wins as companies of all sizes continue their digital transformation efforts and look to move their contact centers to the cloud to better drive omnichannel customer engagement. One of the other announcements we made at Signal was Verified by Twilio, a new feature that allows branded, trusted phone calls that are authenticated by Twilio. This is yet another step that we are taking to take back the voice channel and combat unwanted robocalling. I know that investors have concerns about the impact of robocalls and potential legislation on our business. We've talked about it on previous calls, but given the amount of noise, I wanted to address it head-on today.
When people talk about automated calls or texts, there's really two kinds of behaviors they might be referring to: automated unwanted calls and texts, such as those you might get from a fraudster, and automated but wanted calls and texts, such as those you might get from your child's school. To be clear, we do not support unwanted automated calls or texts on our platform. Since day one, we have had many preventative measures in place, and since the beginning, we have proactively kicked customers off of our platform who have been bad actors. We welcome legislation that will further impinge these bad actors and are working with the FCC, carriers, Congress, and industry groups to support this. That said, we do support customers making automated wanted calls and texts. Think about doctor's appointment reminders, flight changes, fraud alerts, or emergency notifications.
There's a valid concern that legislation or regulation designed to target the bad actors could impact these good use cases. Let me address what we're doing there. As of today, both the House and Senate have passed robocall legislation that we support, and we expect a bill focused on SHAKEN/STIR implementation to be signed in the coming weeks. As a reminder, SHAKEN/STIR uses digital certificates based on common public key cryptography techniques to ensure the calling number of a telephone call is not spoofed. We have made great strides to help ensure calls on our platform will be signed under SHAKEN/STIR and expect to have those pieces in place by the end of the year, in line with the industry goal for initial go live. Just yesterday, we announced that we have joined the board of ATIS, the standards group that developed SHAKEN/STIR.
We are also a member of USTelecom, the industry association primarily responsible for advancing illegal robocall mitigation efforts. Our General Counsel, Karyn Smith, sits on their board. With our involvement at ATIS and USTelecom, Twilio is furthering our commitment to protect voice and messaging channels, partner with carriers and other voice providers to combat illegal robocalls, protect our networks, and ensure we have strong federal policies that benefit consumers and businesses. We continue to anticipate that carriers will proceed cautiously when blocking calls, as they understand that consumers want and need to receive calls and texts from their doctors, schools, teachers, pharmacists, and others. Importantly, since carriers were given the option to automatically block calls back in June, we have not seen an impact on our customers' traffic.
However, even if you looked at the most extreme case of carriers or legislation that allows the blocking of wanted communications, including text messages, we currently estimate, based on historical data, that a single-digit percentage of our total revenue could be impacted. This is obviously the worst-case scenario, but we expect that the actual impact will be de minimis. We feel confident that we are taking the necessary steps to ensure our platform is ready for these changes. Importantly, the FCC has classified text messages as a Title I service, which indicates that the commission intends to keep this channel largely regulation-free. The carriers all support this approach, as do others in the ecosystem, including Twilio. To close, we are still in the early innings of this massive market opportunity, as evidenced by the great customers George is about to talk about.
Our growth rate at this scale and the demand for our platform. We're focused on managing the business for the long run to help every company reinvent how they engage with their customers in this digital era. I want to thank all of our customers for trusting Twilio as their customer engagement platform, and thank all the Twilions around the world for yet another great quarter and for their endless dedication to our customers. With that, over to you, George.
Thanks, Jeff. Our team delivered another strong performance in the third quarter as we continue to build out a go-to-market foundation for the future through to learn and develop the future of communications. We're holding engage events around the world, with stops coming to Washington, D.C., Toronto, London, Paris, and Tel Aviv in the coming weeks. At Signal, we heard from great customers like Netflix, Lyft, Morgan Stanley, and others on why they chose Twilio to power their customer engagement. Let me highlight a few of the exciting deals from the third quarter. We entered into a new relationship this quarter with the U.S. Department of Agriculture's National Finance Center, the shared service provider for financial and HR services for approximately 650,000 employees.
The USDA is using programmable messaging to authenticate their employees via SMS when they are logging in to review pay stubs, W-2s, 401 allocations, and more. In addition to one-time password authentication, the National Finance Center will use SMS to periodically send account notifications and reminders to their employees. This is an exciting new relationship with one of the largest departments of the federal government. We also expanded our relationship with Chime, the fastest-growing U.S. challenger bank platform, focused on helping their members achieve financial health. To accommodate their incredible growth, Chime chose Twilio Studio, our innovative visual application builder, to build an inbound IVR completely from the ground up, serving banking customers for any level of support on their account. They are disrupting the banking industry, and we're excited to support them on their journey.
Last year, we started our expansion in Australia. We're excited by the progress there. We entered into a new relationship with Westpac, Australia's second-largest financial institution and a Global 2000 company with a market value of AUD 100 billion. Westpac is rapidly expanding its digital efforts and needed a new solution for account notifications for its tens of millions of customers. Westpac chose Twilio Notify as its standard across its brands as we offered the most complete omni-channel platform with global scale. We look forward to working together with Westpac as they continue to meet the demands of the modern customer. Flex continues to be top of mind for our customers. We've heard great feedback since we became GA in October 2018. This quarter, we signed a new relationship with Allianz SE, the world's number one insurer and one of the world's largest financial services groups.
Working with a close business and developer approach, Allianz Direct chose Flex to power their customer and contact center agent experience. In combination with our partner Devoteam, Allianz Direct has rolled out Flex in several regions. We also entered into a new relationship with CompuCom Systems, a subsidiary of Office Depot that provides end-to-end managed services, technology, and consulting to enable the digital workplace. As part of their strategy to reimagine their customer experience, CompuCom chose Flex to provide a more seamless customer interaction, regardless of channel, for more than 2,000 agents. These are both great examples of the massive shift in the contact center space, and we are excited to help companies like CompuCom and Allianz Direct adapt to the changing needs of their customers with the fully programmable Flex platform. The Twilio SendGrid cross-sell efforts continue to progress well.
In this quarter, we had a cross-sell win with a major global airline. They've been a Twilio SMS customer for a few years, and in Q3, we expanded our relationship by adding Twilio SendGrid Email, which the company will use for millions of notifications each month, including flight changes, check-in reminders, delays, and more. This was a strong team effort and shows the value of combining messaging and email for customer engagement. I want to personally thank all of the amazing SendGrid team members who are working hard to make deals like this possible and all who built a product capable of supporting these great customers. As we expand our penetration into a market like travel, we're continually encouraged that this opportunity truly spans companies of all shapes and sizes.
On the IoT front, we signed an agreement with Mason, a leading mobile infrastructure-as-a-service provider that Forbes called the Twilio for tablets. Mason was created to help companies build and scale smart hardware products, and its customers have used its services for ordering kiosks, patient engagement in hospitals, and more. They were looking for a partner to scale their edge computing business and turned to Twilio Super SIM to connect the customized tablets they build for their customers. While the IoT market is very early, we're excited about the continued traction of our IoT offering. Overall, we're well-positioned as we head into Q4 and into 2020. We're making the key investments to fuel continued growth going into next year, and we continue executing our strategy to grow our enterprise presence, scale our partner ecosystem, and expand internationally. With that, I'll pass it over to Khozema.
Thanks, George. Good afternoon, everyone. Q3 was another strong quarter for the business. High growth of this scale is a testament to the power of Twilio's platform and showcases how our customers continue to choose Twilio as their trusted customer engagement platform. Diving into the numbers, base revenue, which includes Twilio SendGrid, grew 79% year-over-year to $276 million in the third quarter. Organic base revenue for Twilio was $227 million, growing 47% year-over-year. Twilio SendGrid grew 31% year-over-year to approximately $49 million. As Jeff mentioned, while the growth was very strong, base revenue came in a bit lower than we expected. We've experienced rapid growth, and with that comes some growing pains. Consequently, a few of our older systems sometimes fall short of where we'd like them to be.
We faced some of these growing pains in the third quarter as we discovered some errors in our billing processes that resulted in a few one-time credits being issued to customers totaling approximately $5 million, which will also impact our ongoing run rate. Importantly, our internal controls identified these errors, and we understand the root issues and are working to improve our billing-related processes and other systems. We will continue to invest in systems to support our strong growth trajectory into the future. Let me take a moment to discuss our current revenue disclosures and a change we are planning to make. Given the size and scale Twilio has achieved, we believe the variable revenue designation has become less meaningful and that total revenue is a better way to evaluate the overall business.
Variable revenue has materially declined as a percentage of the total, 7% this quarter, 7% last quarter, and less than 10% in 2018 prior to the closing of the SendGrid acquisition, versus 16% in the quarter before we went public. Accordingly, beginning in Q1 2020, we will no longer break out base and variable. We expect to continue to disclose the contribution from WhatsApp through 2020, which constitutes the majority of the variable revenue category. One important note regarding this change is that Dollar-Based Net Expansion is currently calculated on base revenue. We will be shifting the basis of our expansion metric to total revenue on an ongoing basis, and we will provide historical data using total revenue to normalize your models.
Dollar-Based Net Expansion Rate was 132% in the third quarter, a very strong rate at this size and scale, especially coming off of difficult compares from 2018. The credits I mentioned impacted DBNE by a few points in the quarter. As we approach the end of the year, let me provide a bit of context around what to expect for expansion rate over the next several quarters. First, recall that strong political traffic and the ramp of a large international customer benefited growth by about 10% in Q4 2018. We do not anticipate those to occur again this year, making for a difficult comp in Q4. Second, Twilio SendGrid's historical Dollar-Based Net Expansion Rate had been about 115%, give or take a few points in any given quarter, and will therefore lower the overall expansion rate once combined next year.
As you may recall, SendGrid was acquired on February 1st, 2019, and therefore only contributed two months to our Q1 2019 results. As Q1 2020 will have the benefit of an extra month of revenue from SendGrid, our expansion rate will also benefit, and we expect a sequential increase from Q4 to Q1. The full impact of the combined expansion rate will be realized in Q2 2020. Moving on, we had approximately 172,000 active customer accounts at the end of the quarter. Top 10 active customer accounts contributed 13% of total revenue in Q3, flat sequentially and down from 18% in Q3 2018. Gross margins were at 58.4% in the third quarter, within the mid to high 50s range we've been communicating.
Going forward, we continue to expect to operate in the mid to high 50s while the normal puts and takes, customer, product, country mix, carrier fees, FX, and others may cause minor fluctuations. A2P carrier fees had no impact during the quarter as they have still yet to be implemented. We do not have any updates with regards to timing, but we will provide the P&L impact during the earnings call post-implementation. As anticipated, operating profit was slightly in the red, coming in at a loss of $3.6 million, primarily due to Signal, our annual developer and user conference, which was held in August. Additionally, we refined our process for accumulating inputs that feed our capitalized software process, allowing us to make better estimates. This could potentially reduce the amount we capitalize and more closely align our P&L with cash usage.
As we get into Q4 and fiscal year 2020, we see a tremendous opportunity ahead of us to continue to increase market share and grow the company. George has talked about some key priorities in the go-to-market efforts, and we believe investing in these areas today, as well as our product innovation, will position us to take advantage of this massive market opportunity. In that context, we now expect to finish 2019 with a slight operating loss versus our earlier guidance of $5 million-$8 million of profitability. We believe these investments will drive multiple years of elevated growth. Staying on profitability for a moment, while we're not optimizing for this today, I want to reiterate that this is an important long-term focal point for the company.
We are in the early stages of a number of changes internally to drive efficiencies over time, including changes to strategic sourcing, working capital, and stock-based compensation, to name a few. We've heard your feedback on our stock-based compensation and dilution and are actively working to bring both of those numbers down over time. Some of these levers will also close the delta between GAAP and non-GAAP results. We don't have any specific items to highlight today with regards to these items, but we will provide more information when the time comes. Ultimately, the investments being made today enable Twilio to extend our market-leading position and take full advantage of this generational opportunity. With that, thank you everyone for joining, and I'll now open the call for questions. Operator?
At this time, in order to ask a question, you will need to press star one on your telephone. To withdraw your question, press the pound key. Please stand by while we compile the Q&A roster. Your first question comes from the line of Ittai Kidron from Oppenheimer.
Thanks. Hey, guys. Khozema, I wanted to really focus on the system issue, the $5 million credit, just to make sure I understand this. Is this a cumulative effect over multiple quarters that had to be clear in the quarter, or this is just happened in this quarter itself? Help me get my hands around this.
Yeah, it's a little bit of both. Thanks for the question. As we said on the call, I think, just based on the growth that we've experienced as a company, I think one of the things that we've looked at is that our systems are going through some growing pains, as we explained. This particular issue was related to our billing system, where specifically some order forms were input incorrectly in sort of a manual fashion. Obviously, this is not the experience that we want for our customers, given that customer trust is really paramount. We went back and did a thorough analysis of the problem, and once we found some of these manual inputs, we promptly issued credits to the customers that were impacted.
There were a couple of them that were in the quarter, there were a couple of them that were preceded a couple of quarters. We think we've got a good handle on what caused the errors historically, and we've got controls in place now to address this going forward.
Just to make sure, as I think about the shift from September actual results to December, the September results already take this credit fully into effect, and December is now guided with this in mind. Still, there's very little quarter-over-quarter growth in your core business from September to December. I understand that last year you had that one time, the political, the midterms, and the one-time large customer. It seems like business activity or the volume of transaction is still decelerating quite substantially from September to December. Is there anything else in here that you think requires highlighting in understanding this?
Yeah, I wouldn't characterize it that way. Let me give you a little bit of the financials, and then I'll have Jeff give you a little bit more of the detail, and George as well, maybe from a customer perspective. I think in terms of the December quarter, we did take our guidance down by a bit more than the ongoing run rate. I think we feel like we've got a good handle on the issues, and we're certainly actively working to correct them. I think we felt it was warranted to put in a little bit of an extra buffer until we're absolutely certain that all the issues are addressed. We're not necessarily quantifying that, but I think we felt like it was prudent to do so as we work through the fixes that are entailed with that.
As it relates to the year-over-year, any way that you look at it, really, we feel like that we've got terrific growth at scale, that it's really impressive. The comps, as you mentioned, we've signaled for a long time were going to be difficult in the back half of the year. You pointed to the political traffic. I think we reiterated again, for example, that we had a one-time customer that ticked up, particularly last year. I think we've been pretty clear that around our expectations that the growth rate and the expansion rate would slow down a little bit in the back half of the year. Just maintaining these rates is obviously hard at this size.
Again, the credits that we referenced, those impact a little bit our base revenue growth, as I explained a moment ago, and will have a little bit of an impact on DBNE too. We still feel great about where the business is headed, and let me turn it over to Jeff, maybe to add a little bit more color.
Hey, Ittai. Despite the manual errors, I think we turned in a quarter growing 47% year-over-year organic at nearly a billion-dollar run rate. We're really proud of what we've accomplished. We're one of the only companies growing at this scale at this rate, and we're very proud of that. We're focused on the long-term opportunity. This is one of those once-in-a-lifetime changes that's happening in how companies engage their customers using all these new digital channels, and we're here to capitalize on that mega trend. We're hearing all the right things from our customers about what they need and our ability to provide it. I look at it's a great market, we've got a great business model, we've got a great team, and I see them all getting better as we grow. I'm really excited about what lies ahead.
Very good. Good luck, guys.
Thank you, Ty.
Your next question comes from the line of Bhavan Suri from William Blair.
Hey, guys. Thanks for taking my question, thanks for the billing clarity there, Khozema. I guess I did want to touch on the Dollar-Based Net Expansion Rate here. Even if we add a few hundred basis points from the credits, the billing issue, it feels like it decelerated reasonably quickly. I guess I'd just like to understand sort of, obviously scale matters, but it wasn't a basis point or 100 basis points or 200 basis points. I was just trying to understand if you could provide some color on the dynamics that have affected this metric. I guess, are you seeing changing spending habits within customers, was it just the volatility of how applications have launched that drives it up and down, the speed of that? I guess just some of that color would be helpful.
Yeah. Let me provide a little bit of the math, and then maybe I'll have George comment from a customer perspective in terms of what he's seeing. I guess just as a starting point, I would say the short answer is no. Maybe just to reiterate, Bhavan, I think from our perspective, the growth at the scale that we've reached is really strong. I think, again, we always knew that we would have pretty difficult comps in the back half of the year. If you recall Q3 2018, we had growth that was approaching 70%, and then going into Q4, we had growth that was approaching 80%.
We had a little bit of political traffic in there that was elevating that, and then we had this one-time customer that we actually tried to back out of the results last year to try to normalize it a little bit. That had an impact as well on the expansion rates back in that period. As we lap the year, I think we've been pretty clear in our expectations that DBNE would, in fact, fade over time, in part because of the comps, in part because the business is just becoming a lot larger and we're running into the law of large numbers. Then, the footfall in this, if you will, is this manual error issue that we elaborated on as it relates to billing associated credits.
You put all that together, and that's what kind of results in the math that you saw on the DBNE that we got in the quarter. I don't think it's anything related to some of the issues that you pointed to. Let me turn it over to George, and he can color on that a bit more.
Sure. Thanks, Khozema. I would say that we don't see any fundamental change in our demand environment. We still see a lot of interest from customers, and we haven't fundamentally seen a shift except for some of the dynamics that Khozema talked about, which are nothing to do with really our customers. We're excited, we're investing, and we still plan to capture the big opportunity in front of us.
Okay, helpful. Let me quickly just ask a follow-up on Flex here. Obviously, nice wins there. I guess maybe, George, for you, maybe touch on what sort of demographics are you seeing from the early adopters, and then how are the initial conversations going with the VAR channel? This is something you and I have talked about multiple times in the past. Just love to get color on those two. Thank you.
Yeah, great question. In terms of the demographics of the customers, I would say generally speaking, it's consistent with what we've said before. It's still a newer product. We're still mostly in the early adopter phase. We have a lot of interest from digital economy customers and very tech-forward companies. We're excited now with Allianz and CompuCom to start to see the beginnings of expanding that more towards the mainstream. I would caution that this is not the normal pattern, or we haven't yet crossed the chasm. It's still a newer product. We're excited about these green shoots, and I think it speaks to the power of the value proposition of the product.
In terms of the partners, we're excited about the traction we're seeing there, and, as I've mentioned, I believe in the past, we've not really historically at Twilio had a strong reseller motion or, frankly, infrastructures for resellers. We've been investing in that this year, and are bringing some initial resellers onto that platform. We're testing it right now. I think it'll be something that we'll be able to onboard more in 2021. I don't think it'll be material for a while for us, but we are investing in it, and certainly, I think Flex is definitely the catalyst that is growing interest in that for us.
Great. Thank you, guys.
Thanks, Bhavan.
Thanks, Bhavan.
Your next question comes from the line of Mark Murphy from J.P. Morgan.
Yes, thank you. Khozema, the $5 million billing impact to Q3, is it also $5 million in Q4? I'm just wondering for how many quarters do you think that would continue into 2020?
Thanks for the question, Mark. Good question. I don't think we see it necessarily continuing into 2020 in the way that your question intimated. Obviously, it'll impact the way that we evaluate our run rate from an FP&A perspective. In general, we feel pretty good about how things look after Q4. We're obviously not guiding to that today. I don't want to get too far down the road on that one. As it relates to Q4 itself and the way that we characterized our guide, as I said a moment ago, it definitely does have an impact on the ongoing run rate. It's about a full $5 million.
I think as you probably understood based on the comments that we gave a moment ago, we have put a little bit of a buffer, if you will, into the way that we've guided into Q4, just until we have our arms fully around the issue. We just felt like that was the smart thing to do in terms of the way that we communicated the way that things would play out for investors. While we feel like we have a good handle on the issue, we do want to get through all the fixes. It's kind of a one-time thing in nature. We feel we caught it ourselves, and we think we're more or less out of the woods, but we want to be smart about the way that we convey it.
Okay, great. How many customers will be receiving the credits?
I don't have that number off the top of my head, Mark. It's pretty small. We're talking about a single-digit number. I don't know it off the top of my head, but it's in that realm.
Okay. The other question I had for you, I think we're looking at base revenue growth. If we look at it sequentially, it's around 7% in Q3, and then you're guiding about 9% for Q4. If we trended that out and just considered how that would look across four quarters, I think it would get us to something like low 30s to mid-30s. I'm just curious, would you view that as a fair assessment of the glide path on the base revenue currently and maybe how we'd trend it into next year? Would you look at this and say it's artificially constrained, I guess, again, due to the billing errors and what sounds like a little extra cushion for Q4?
Yeah, Mark, I'm just not going to comment on anything into 2020. We'll guide to that obviously when we do the Q4 call. I would simply say that in the back half of this year, we have some tough comps as we've signaled previously. We obviously had the political, we had that one-time customer that we've alluded to previously, we've got this billings thing.
You're going to have to kind of put that together in your model, but I think we're just not in a position to be able to guide out to beyond 2020. What I will say is that we feel great about what the results were in the current quarter, 47% organic in spite of the comp, in spite of whatever the self-inflicted issues, and we feel great about the overall inputs to the business.
Thank you.
Your next question comes from the line of Brent Bracelin from Piper Sandler.
Thank you, and good afternoon. One for Khozema and one for Jeff, if I could. Khozema, I wanted to maybe take a little different slice at the slowdown you saw here. If I just look at the top 10 customer contribution, versus the revenue outside of the top 10, it looked like the bulk of the slowdown, and in fact, you had a $11 million sequential decline in revenue at the top 10. Growth outside of the top 10 actually accelerated from about an 84% growth to 85% growth in the quarter. As you think about the cohort of top 10 customers that did decline, is it safe to assume that that's where some of the bulk of the credits were applied? Maybe could you just talk about net expansion rates looking at this cohort outside the top 10 that actually accelerated in the quarter?
Is there broad-based change to Net Expansion Rates, or do you think some of these top 10 customer declines sequentially are having a bigger impact on just that calculated Net Expansion Rate? Again, one quick follow-up for Jeff.
A lot in that particular question. Just to start off with, in terms of the credit itself and where that was placed, or the credits, as I said earlier, there were a handful of them, single digits. There were two different customers that would've been in our larger customer base. Off the top of my head, I don't think they were in the top 10, maybe top 20, but among our larger customers, to be sure. In terms of the overall dollar value generated from that top cohort of customers, we're pulling that up here, I think overall, that was up actually on a quarter-to-quarter sequential basis. Feel good about obviously the fact that the % of total was about flat, which speaks to the concentration of our overall top 10, kind of holding steady or declining as time goes on.
I think the premise of the question is incorrect. We'll follow up on the exact math with you offline or in the callbacks. I don't think that's right.
Got it. I thought it was 18% last quarter, 13% top 10 mix this quarter. It was 13 last quarter, 13 this quarter?
That's right.
Okay. Then Jeff, for you, getting some questions around RCS. Obviously, you saw 4 months ago, Google going kind of direct with RCS. You have 4 months later, that was in Europe, 4 months later now, you have Verizon, Sprint, AT&T, and T-Mobile joining up. I guess the question we're getting from investors here is how should we think about the potential impact of pricing, relative to kind of this next gen SMS RCS service, this new consortium, and what potential impact that could have on the overall business as you think about the next couple of years? Thanks.
Yeah, absolutely, Brent. Thanks for the question about RCS. First of all, Twilio has supported RCS since early 2018, and we're a strong supporter of RCS as a channel. Just as a reminder, you alluded to this, but think of RCS as an upgrade to SMS, right? The reach is similar, the economics are similar, but it adds new capabilities that you can do within the message. Things like commerce or maps or call to action buttons and more. This is really exciting because we and the developers building on Twilio can build even more rich experiences inside of the core messaging app that's already on all the Android phones. That's really exciting. It opens up new opportunities.
Now, the CCMI, the Cross Carrier Messaging Initiative that was announced by the four carriers in the U.S. this week, is about building interop between those four carriers, which we view will hasten its implementation and its rollout, which is good, so that we and our customers can really start leveraging it because the promise of RCS has been out there for a long time. It's just all been dependent on carriers rolling it out, and when you get the top four carriers in the U.S. agreeing on rolling it out and how they're going to do that and interoperating with each other, that's good. It means that our customers can get the benefit of RCS probably sooner than if the four carriers weren't interoping and cooperating on this effort.
Your next question comes from the line of Nikolay Belyaev.
Hi, thanks for taking my questions. My first question is for both George and Khozema. When you look at the expansion rate and specifically the customer cohorts outside of the top 10, what are you seeing? Are the newer cohorts trending in line with the older cohorts? Part B of my question is, are you happy with the expansion rate of the older cohorts?
Hi, this is George. Yeah, for our newer cohorts, I wouldn't say that the expansion rate characteristics of them have changed materially from quarter-over-quarter.
Okay. I guess, a different question for you both. Sales and marketing as % of revenue has ticked up for the first time above 25% for a while, versus historically 20%-25% of revenues. Is 25%-30% kind of like the new normal in terms of sales and marketing spend as you guys overlay a direct sales force?
Yeah, this is Khozema. I wouldn't say per se that it's the new normal. I think what we've said for a while is that we feel like that we want to continue making investments in the sales and marketing engine, so long as we believe those investments are efficient. I think the one particular outlier as you look at the, excuse me, Q3 numbers is Signal, which shows up as a large one-time expense and also is a driver for the loss that we took in the quarter. I think you're comparing a little bit of an apple and an orange there in terms of Q2 to Q3.
Got it. Thank you.
Yeah. Thanks.
Your next question comes from Alex Zukin with RBC Capital Markets.
Hey, guys. Thanks for taking the question. Maybe just on the favorite topic of Dollar-Based Net Expansion again. In terms of the guidance for Q4, if we add, and I know this has been asked a little bit, but I'm going to try it a different way. If you add back the credits, you do still see a fairly material drop-off that I think, by our math which, correct me if I'm wrong, looks like it's in the very low 120s for Q4. You mentioned both the extra cushion and the bounce back in Q1, in terms of Dollar-Based Net Expansion, and then completing the anniversary-ing. Is there any way you can help us just understand which ZIP code should we be thinking about for a sustainable Dollar-Based Net Retention? Is it in the 120s? Is it in the teens? Is it higher than that?
Just any help there would be appreciated. I have a quick follow-up.
Yeah. This is Khozema. Thanks for the question. I'll give you a little bit of a directional math. We don't guide on DBNE specifically. I'll try to help you in terms of updates to your model. I think the way to think about it is this: is that in Q3 and Q4 both, as I mentioned earlier we had extremely elevated growth rates in Q3 2018 and Q4 2018. Approaching 70% in Q3 and approaching 80% in Q4. As a result, you had this one-time political dynamic, obviously, which contributed pretty significantly. You also had this one-time customer that we called out. Those impacts are obviously stripped out in the back half of this year in the absence of an election and with that customer behavior having normalized.
The second dynamic is that, as you pointed out, you do have this credits issue, which also affects a little bit our revenue run rate into Q4 as well. You do have a little bit of a fade in both Q3 and Q4 relative to prior periods in terms of that overall DBNE. Again, we don't guide on that number specifically. I think you're a little low, based on the way that you marked it, but I don't want to provide any real specificity beyond that. As we get into Q1, it becomes a little bit more complicated, obviously, because you've got a January stub month, and then you've got the anniversary of the SendGrid acquisition, at which point we'll start to fold in that DBNE with our overall DBNE, which just by definition is going to make it come down a little bit.
Based on the way that we're currently calculating it, what I will say is that we don't expect that combined rate to come down as much as I'd previously suggested. Again, we're not providing guidance around it for Q4 or out into 2020, but we feel quite good about continued expansion out into 2020.
Got it. Maybe just a follow-up for Jeff or George on Flex adoption. You signed a nice customer in Allianz. I think you guys have previously talked about this business potentially being a pretty meaningful contributor to overall revenue over the next few years. I'm just curious, how should we think about this business over the next two years from a contribution to revenue perspective on a decent timeframe?
This is George. Without commenting specifically on revenue timeframes, I'll let Khozema take that on. Your question on adoption, we are excited about the adoption we're seeing of our customers. I think that obviously people that are interested in buying Flex are obviously talking to our existing customers, and I think that's a sign that we are seeing success with existing customers in the field. Yeah, we're encouraged by certainly these data points to kind of continue us on our journey. As I said, it's a long journey. We're early, but with a huge market opportunity, we think we're a very differentiated value proposition. Yeah, we definitely think that this is something that has potential for the long term to be meaningful.
In terms of its contribution to revenue, this is Khozema. I think it's in our revenue today, obviously. I mean, we generate revenue off of the product today, and obviously that will continue to grow as time goes on. It's not really a material contributor to revenue yet. We certainly expect it to be over time. I think what I've signaled in the past is that our expectation is certainly that that would happen over the next several years, and we expect that to be a fairly material number as time goes. I think in terms of its material contribution to revenue, I wouldn't really anticipate that until really the back end of 2021, maybe 2022. As George said, we feel great about where things are headed and some of the most recent customer wins.
Perfect. Thank you, guys.
Thank you.
The next question comes from Heather Bellini with Goldman Sachs.
Hi, this is Dan Church on for Heather Bellini. I just have a quick question. It's been about nine months since the SendGrid acquisition. Any color commentary you could give us in terms of how cross-selling initiatives have tracked relative to expectations?
Yeah. Hi, this is George. Yeah, I think you heard our example of the airline that we closed in Q3. I would say generally speaking, we're on track. I think we've fully integrated, at this point, the sales forces, and we are going together to existing customers and new customers to cross-sell the products. Yeah, I think I would characterize them as on track.
Helpful. Thanks. I think the rest of it might've been asked. Thank you.
The next question comes from the line of Meta Marshall with Morgan Stanley.
Great, thanks. I wanted to refer to kind of the expansion agreements you alluded to in the prepared script, just to get a sense of timeline of those contracts closed or perhaps more broadly, like as you head into applications, what you're seeing around sales cycles when you're not dealing directly with the developer. Then maybe on a second question, gross margins obviously remained very healthy in the quarter. Is there any customer or account churn we should be kind of thinking of from pricing competition or deals that you kind of walked away from? Thanks.
This is George. I would say that, in terms of the sales cycles, I think you asked a good question in the sense that we definitely do see 2 types of sales cycles. I think the ones where we have a developer self-service start in the account definitely can be relatively shorter cycles. In our commercial segment, those can be intra-quarter transactions at times, so start and complete within the quarter. As we've said, for example, on the other side of the spectrum, for Flex, which typically is not going to be started by a self-service developer start, it's more of a sold product, so to speak. We've been saying now that for enterprises, these are going to have longer sales cycles. I think we're seeing that to be true.
They're having what I would describe to be more normalized enterprise cycles, and I think that's frankly to be expected, and I think it's totally fine. Can you remind me of the second half of your question again?
Gross margins.
Gross margins. In terms of customers leaving, yeah, we haven't seen a material change in churn or logo churn, customers leaving, from that perspective. I don't know if Khozema wants to provide any more gross margin color, but that's what we're seeing in the field.
Yeah. As it relates to churn, I think George answered the question. I think the only other thing I would say about gross margins is it's in the zone that we have been signaling for a long time. It's kind of where we expected it to be. The only thing to really call out is that it was clipped by about 60 basis points due to the billing issue that we've talked about a lot during the course of this conversation. Otherwise, it moves around a little bit within that zone due to a variety of factors, and we've talked about those in the past.
Great. Thanks.
Your next question comes from the line of Nandan Amladi with Guggenheim Partners.
Hi, thanks for taking my question. As you merge the basic and variable revenue streams, what may be a way to think about leading metrics as we head into next year?
Well, I think, maybe just to take a step back for a second. I think the reason that we're doing that in the first place, and I think that variable revenue designation has become a lot less meaningful, certainly in terms of the way that we manage the company internally, the way that we run the company. I think it was a change that we felt like we should make. It was certainly something that I've been thinking about a long time and since I joined the company. Kind of wanted to see the way that it would play out this year in terms of that particular metric. It's also declined pretty substantially, 7% of revenue now. I think what you'll see from us going forward around growth is simply that total revenue, instead of necessarily breaking out base or variable.
We talked about the fact that we would continue to provide some data around WhatsApp, given what a significant percentage of variable revenue that's been in the past. Otherwise, I think we feel pretty comfortable around total revenue as the metric. We'll obviously report on things like DBNE and ARPU, et cetera, otherwise, that's the suite of metrics that you should expect from us.
Thank you.
The next question comes from the line of William Power with Baird.
Great, thanks. Yeah, Khozema, may just come back just quickly to the Dollar-Based Net Expansion Rate. I guess the one other piece that maybe I missed, I think you said you pulled the variable revenue into that calculation as well. Is that going to begin in 2020 that you do that? What's the anticipated impact just in terms of kind of level setting in front of that from, again, WhatsApp and the variable revenue being part of that?
Hi, Will. You mean on DBNE specifically, what's the impact associated with folding in variable as part of the total calc?
Yes.
Yeah. I think it's going to be pretty de minimis, honestly. I think that's one of the many reasons that we've been thinking about doing this for some time. I think you won't see a material impact. There could be an impact, plus or minus 100 basis points. I don't really think it'll be material and something for you guys to think about. Okay. There was our second part of the question you missed.
Yeah, no, that was it. I just wondered.
Yeah, the timing of the change. Yeah, sorry. The timing of the change is we'll make that change in Q1. We'll provide you all with a bridge, to make sure that you can see the different moving pieces so that we're not just bringing a number on you guys for the first time. We'll kind of walk it through.
Okay. Then maybe just a question, for Jeff or George, just the Conversations API seems like a lot of excitement at Signal from that. I recognize in beta here, but any further color you can provide on early use cases, customer feedback, how you're thinking about the excitement you're seeing as that kind of moves into 2020?
Absolutely, Will, this is Jeff. Just a quick background. We launched Conversations to help companies to connect better with their customers, right? We've all had that experience. A few years ago, it was kind of novel when you got a text message saying your flight was delayed or your flight was boarding or your package got delivered, and that was exciting. When you'd reply to the message, it was like nothing would happen. You'd get back an automated thing that said, "If you need help, call us." It's just not a great experience.
What we're seeing the leading companies out there do is say, "Hey, look, obviously there's a lot of alerts and things like that, and then you should be able to reply, and the company should be able to route it to the employee who can best help you." Sometimes it's an in-store employee. Sometimes it's a contact center employee. It's a variety of workers who instead of, like 20 years ago, all these people may have been sitting at desks with a desk phone, and now people are out and about. They've got a BYOD device in their pocket, and messaging is the preferred medium. Really the way consumers connect with the employees of a company has changed.
Conversations is the product designed to capture that change and enable every company to be able to service their customers in the way that those customers really want to talk to the company. I'll say, it's a brand-new product, obviously. We just announced it at Signal, so it's a beta product. It's the very earliest stages of that product. I think there's a macro trend here going on that starts, as often happens in the consumer world, with how we engage with each other and then bridges into how we want to engage with companies. That notion that I want to text with the company is not necessarily new. It's just in the past, it's been a one-way thing for most companies. With Conversations, we're trying to answer the call of making that a two-way dialogue.
What I'll say is that since we announced Twilio Conversations, no pun intended, it's created a lot of conversations with our customers. I think we've struck a nerve there, and it's started to pique interest from customers. It's super early in the life cycle of that product.
Yep. Thank you.
Your last question comes from the line of Richard Valera with Needham & Company.
Follow-up question on SendGrid. Doing the math, it seems like it grew just over 30%, in the quarter, which is, I think, a little better than it's been growing recently and well above, I guess, the 25% sort of bar you'd initially set. Just wondering, given that you're still, I think, somewhat early in the cross-sell process, where you think the growth rate for that could go, and could we see that sustained at a 30%-plus growth rate in next year as you kind of really flesh out the cross-selling initiative?
Yeah, Rich. Hi, this is Khozema. In terms of the growth rate for the quarter, it came in at 31%, so your math is right, just above 30 at 31. We feel pretty good about the growth prospects of the business. We're not going to guide out into next year, but I think as I've said previously, we would be disappointed, I think, if the growth rate fell below that, and I think we have a really great shot at it being above that. That's kind of in the zone of our expectations. I don't want to provide any detailed guidance around the number, but our expectations have consistently been somewhere in that range.
That's great. Just one follow-up, quick on the expense side. Should we think of the 4Q expense levels as sort of the run rates for going forward and sort of scale upward from them as we head into next year? Is there anything exceptional in Q4 that we would back out or is it sort of a good run rate to serve as a base going into next year?
Yeah, that's a good question. I think we're still figuring out what the 2020 plan is going to be, right? I wouldn't one way or the other read anything into Q4 necessarily. What I will say is this, is that I think the way that we think about it is that so long as we see multi-year elevated growth at scale, we certainly see great opportunities to put investable dollars. There's an opportunity for us to continue investing and continue generating great returns for investors. There's nothing one time per se. Otherwise, we're looking at our investment deck for 2020 as we speak, and those are the kinds of returns that we're expecting is kind of what you saw in the Q3 from us, 47% growth organic at scale, and hopefully we can deliver something elevated in the future.
Okay. Thank you.
There are no further questions at this time. Thank you for participating in today's conference. You may now disconnect. Goodbye.