Rubico Inc (RUBI) 2017 Q3 法說會逐字稿

Good day, everyone, and welcome to the Rubicon Project's Third Quarter 2017 Earnings Conference Call. (Operator Instructions) Please note that today's event is being recorded. I would now like to turn the conference over to Nick Kormeluk. Please go ahead.

Thank you, operator, and good afternoon, everyone. Welcome to Rubicon Project's Third Quarter 2017 Earnings Conference Call.

As a reminder, this conference call is being recorded. Joining me on the call today are Michael Barrett, President and CEO; David Day, our CFO. I would like to point out that we have posted slides to accompany today's presentation on our website.

Before we get started, I would like to remind you that our prepared remarks and answers to questions will include information that might be considered to be forward-looking statements, including but not limited to statements concerning our anticipated financial performance, including without limitation, revenue, advertising spend, profitability, net income or loss, adjusted EBITDA, earnings per share and cash flow; strategic objectives, including focus on header bidding, mobile video orders and private marketplace opportunities; investments in our business; development of our technology; introduction of new offerings; the impact of our acquisition of nToggle and its traffic shaping technology on our business; scope and duration of client relationships; our fees; business mix; sales growth; client utilization of our offerings; our competitive differentiation; our leadership position in the industry; our market share, market conditions, trends and opportunities; user reach; certain statements regarding future operational performance measures including ad requests, fill rate, advertising spend, take rate, paid impressions and average CPM; and factors that could affect these and other aspects of our business.

These statements are not guarantees of future performance; they reflect our current views with respect to future events and are based on assumptions and estimates that subject and -- are subject to known and unknown risks, uncertainties and other factors that may cause our actual results, performance or achievements to be materially different from expectations or results projected or implied by forward-looking statements.

A discussion of these and other risks, uncertainties and assumptions is set forth in the company's periodic reports filed with the SEC, including our 2016 annual report on Form 10-K and quarterly reports on Form 10-Q for the 3 quarters of 2017 under the headings Risk Factors and Management Discussion and Analysis of Financial Conditions and Results of Operations. We undertake no obligation to update forward-looking statements or relevant risks.

Our commentary today will include non-GAAP financial measures. Reconciliations between GAAP and non-GAAP metrics for our reported results can be found in our earnings press release and in the financial highlight deck, which we have posted to the Investor Relations website at rubiconproject.com.

At times, in response to your questions, we may offer incremental metrics to provide greater insights into the dynamics of our business. Please be advised that this additional detail may be onetime in nature, and we may or may not provide an update on the future of these metrics.

I encourage you to visit our Investor Relations website to access our press release, financial highlights deck, periodic SEC reports and webcast replay of today's call to learn more about Rubicon Project.

I will now turn the call over to you, Michael. Please go ahead.

Thank you, Nick, and good afternoon, everyone. Welcome to our Q3 2017 earnings call. When I arrived in March, I focused the team on accelerating our efforts on header bidding. At that time, it was the greatest and most immediate opportunity that we had to win back market share. And today, I'm happy to say that the team has made significant progress against this objective in just 6 short months.

The amount of inventory that we have access to has grown by 50% year-over-year. According to a recent report issued by Rockford, within Prebid open source client side header bidding implementations, we are now one of the top 2 revenue sources for desktop inventory being sold through these implementations. And in mobile ad spend, we were neck and neck with Facebook's Audience Network as the leader.

And with the launch of our open source, server-side header bidding solution, we expect to further grow our market share and strategic position within header bidding as sellers move towards the use of open source, server-side solutions to meet their monetization needs in a rapidly evolving market.

Now that we are no longer playing catch-up in header bidding, we are shifting our efforts to focus on differentiating our platform. In alignment with our strategy, which I outlined on the earnings call last quarter, I see 4 areas of differentiation: first, cost. We are committed to becoming the low cost provider; second, transparency and choice. Clear, fair and transparent auction dynamics that support multiple versions of header bidding implementation; third, the programmatic supermarket. If it can be bought or sold programmatically, you can find it on the Rubicon Project exchange; and fourth, inventory quality and brand safety, which have always been key differentiators for our exchange and an area that we plan to continue to leverage.

Let's start with cost. Earlier this year, we proactively made cuts to our pricing that lowered our total take rate from 24% in Q4 2016 to what's now just under 17% at the end of Q3.

We did this to be more competitive and have seen improvements to our win rates from these changes. However, to win a greater share of ad spend from buyers and to provide the transparency they want, we simplified our pricing model and completely eliminated our buyer fees across our platform as of November 1. This means that buyers will find simplicity and clarity working with us, and their budgets will now go further in reaching the audiences that are most important to them.

As evidenced through direct feedback and public commentary, buyers want more transparency into where their dollars are being spent, and where their ads are appearing. They no longer want or need to work with dozens and sometimes more than a 100 different exchanges to reach the audiences that they want.

Buyers want to consolidate their spend on the platforms that deliver the best results. And we believe that Rubicon Project is strategically positioned to capitalize on this new trend due to the elimination of our buyer fees.

The elimination of our buyer fees drives lower overall take rates than we first envisioned, resulting in greater near-term operating losses. But it's the absolute right action to position us for the future. In fact, we are already receiving positive feedback from key buyers on our exchange, including MediaMath, who told us that they support our decision to enact a simpler, more transparent fee structure as it further enables their clients to efficiently purchase impressions and audiences that are the most valuable to them, while gaining greater visibility into the supply chain.

The heart of transparency at Rubicon Project is delivering a clean, fair, efficient and transparent marketplace, so our clients can optimize their businesses to achieve the best results.

Rubicon Project has traditionally only offered a modified second prize auction within its open and private marketplaces. However, since the introduction of header bidding, there have been a variety of auction dynamics being used by different exchanges and supply side platforms, which directly impacted the ability of buyers to win valuable impressions as their bidding strategies were optimized for only second-prize auctions.

As an exchange, our goal is to facilitate the transaction of inventory at a price that is competitive for our sellers, but still aligned with the value we will deliver to our buyers. To that end, we have been assessing the impact of watching dynamics in our marketplaces.

In the third quarter, working closely with our demand and supply partners, we initiated more comprehensive testing that examines the effects of modified first and modified second-price auction dynamics in our current environment.

This testing was initiated on all header bidding traffic on our exchange in early September. We confirmed that the change of dynamics paired with lower marketplace fees, and in some cases no buyer fee, resulted in better results for both our buyers and sellers.

For instance, the amount of revenue that a seller achieves through our header bidding solution has increased steadily since the initiation of testing. And we have also seen a positive upward trend in the win rates for our buyers as well as an increase in ad spend. These results were achieved before the deployment of nToggle's technology across all of our U.S. buyers.

However, upon the integration completion, which remains on track for Q4 2017, we expect to see even better results related to buyer win rates, publisher revenue and ad spend increases in the future.

We will share more around the results of our auction dynamic tests as we have them. Now that I've discussed the steps we are taking to become the lowest-cost and most transparent platform buyers and sellers, I'd like to shift focus and discuss the progress that has been made in becoming a high-volume global ad exchange. We've achieved great progress towards our goal of being the supermarket of programmatic inventory. If it can be bought or sold programmatically, you can find it on the Rubicon Project exchange.

Today, buyers bid on every type of ad inventory, including desktop, mobile web, mobile in-app, audio, video, digital out-of-home and more. And sellers can reach the premium brands and advertisers that they want associated with their sites and applications. By continuing to grow the amount and variety of inventory types on our platform, we believe that over time, buyers will migrate off of specialized smaller platforms and increasingly conduct business with scaled one-stop platforms like ours.

This quarter, mobile ad requests have increased by 95% year-over-year. And mobile now accounts for nearly half of our ad spend. One of the largest areas of growth within mobile, and mobile app, specifically, is audio. We've seen a 3x increase in ad requests for audio, one of the emerging media channels that we've been investing in. This growth is due to the increasing global user base of streaming audio platforms, like Spotify.

Our video business continues to outpace market growth, with ad spend up 73% year-over-year as compared to 26% year-over-year growth for the industry according to a recent Nielsen report, with video ad request growing more than 60% year-over-year in Q3. And today, I'm happy to announce that there are more than 20,000 buyers accessing video inventory through the Rubicon Project exchange.

And last but not least is orders. We've seen continued growth in orders with the most meaningful contribution coming from private marketplace -- marketplaces, also known as PMPs. We've seen increasing demand from leading global agencies and holding companies that are looking to reach their key target audiences in a brand safe environment. And our PMP technology enables them to strike deals for specific inventory and audiences directly with publishers and ad developers.

We've also made strides to make access to these Private Marketplace packages as easy and accessible to buyers as possible through new partnerships with leading DSPs. Most recently, we signed a partnership agreement with Google to make our PMP packages available to buyers through the DoubleClick Bid Manager marketplace, the first partnership of its kind for Google's DBM marketplace.

Partnerships like this offer a unique value for our sellers, as it makes our inventory packages available to more buyers.

The last key differentiator that I'd like to discuss is one that, as I previously noted, has been a unique selling point for Rubicon Project for a very long time: Inventory quality and brand safety. This was an area that was invested in during the early days of Rubicon Project. And we remain vigilant in constantly evolving our processes and technology for identifying and blocking both bad ads and bad actors from entering our exchange.

Our process is unique, as it requires human oversight in addition to cutting-edge technology. This is an area that we intend to focus on in order to deliver a best-in-class experience and results for both our buyers and sellers.

The investments that we've made in this area position us well for new industry initiatives, such as ads.txt.

There has been a lot of change in the industry. However, nothing has changed in terms of Rubicon Project's strength of offering and strategic position. As with any exchange, our growth is linked to our ability to drive positive network effects. By powering an exchange that is as efficient and cost-effective as possible for buyers and sellers, we are able to better -- we're better able to capitalize on positive network effects and further increase our market share.

We strongly believe that we can and will win in the new ecosystem and in the long run, if we match our superior tech offering and client service with the right pricing for the market.

With that, I will hand the call over to David to give more detail on our financial performance in Q3 2017.

Thanks, Michael. As Michael highlighted, we continued to take very specific actions in the third quarter to execute against our strategy to make Rubicon Project the most attractive platform for buyers and sellers to trade on for their programmatic advertising needs.

The long-term positive outlook for programmatic digital advertising growth makes our space compelling. And we are well positioned in the most attractive growth areas of programmatic. However, getting back to growth will require investment and operating losses -- as evidenced by our third quarter losses, which we expect will further increase in the fourth quarter due to the elimination of all our buyer fees and much lower corresponding take rates.

Turning to our results. For the third quarter of 2017, we generated $195 million in advertising spend, $35.2 million in net revenue; adjusted EBITDA loss of $2.3 million, a loss of $0.14 per share in non-GAAP EPS.

Total advertising spend decreased 5% in Q3 versus Q2 of 2017 and declined by 20% year-over-year, driven primarily by desktop ad spend, which was lower than prior year by 60% and by lower fees across our platform. This is in line with our prior comments that ad spend has begun to stabilize at levels experienced in the first half of 2017.

Total mobile ad spend increased nicely, with a 7% improvement over Q2 2017 and a 10% increase year-over-year, driven by mobile app.

Mobile app ad spend, which is an area of focus for us, continued to grow very nicely and is now over half our total mobile business. Ad spend mix for the third quarter was 47% mobile and 53% desktop, representing a continuing trend towards mobile from Q2 of 2017, when mobile represented 42% of total ad spend. This also represents a sizeable increase over Q3 of 2016, when mobile made up 34% of total ad spend.

Non-GAAP net revenue for the third quarter declined 18% sequentially from Q2 2017, primarily due to a decrease in take rate and secondarily from slightly lower ad spend. Non-GAAP net revenue declined 42% year-over-year, primarily due to a decrease in take rate, ad spend and our exit from the intent marketing business in Q1 of 2017.

Our take rate was 18.1% in the third quarter of 2017, a decrease of 290 basis points sequentially and 680 basis points on a year-over-year basis. Our Q3 2017 exit take rate was 16.9% and will continue to decline in Q4 due to the removal of all of our buy-side fees effective November 1.

Take rate is defined as non-GAAP net revenue divided by total ad spend. The year-over-year decline in take rate was primarily due to our price reductions made to improve our value proposition and grow share in a more competitive environment due in large part to header bidding.

The sequential decrease in take rates in Q3 is largely related to the specific pricing actions we took, as discussed last quarter. In addition, mix shift also had an impact on take rates. For example, the shift to header bidding and a shift from RTB to orders with header bidding and orders carrying -- both carrying lower take rates.

Operating expenses for the third quarter of 2017 were $51 million, excluding a goodwill impairment charge, down from $68 million in the same period a year ago, representing a decrease of 25%.

The impairment charge taken was for all of the remaining $90.3 million in goodwill on the balance sheet, including goodwill from our acquisition of nToggle during the third quarter.

The reduced near-term cash flow outlook as a result of lower take rates and the decision to eliminate our buyer transaction fees, combined with the lower company share price and market capitalization, were drivers of the impairment charge.

On an adjusted EBITDA basis, operating expenses for the third quarter were $37.5 million, down from $45.3 million in the same period a year ago, reflecting the impact of our cost-cutting initiatives from late 2016 and earlier this year.

Net loss was $103.6 million in the third quarter of 2017 as compared to net income of $3.5 million in the third quarter of 2016. The change in net loss year-over-year was due to the $90.3 million goodwill impairment charge and due to lower GAAP revenue, partially offset by reduced operating expenses and a lower net tax benefit for the quarter.

Adjusted EBITDA loss was $2.3 million in the third quarter of 2017 as compared to positive adjusted EBITDA of $15.3 million reported in the same period 1 year ago. The decrease in adjusted EBITDA was driven primarily by a decrease in non-GAAP net revenue partially offset by lower cost, as previously discussed.

Diluted GAAP loss per share was $2.11 for the third quarter of 2017 compared to diluted GAAP earnings per share of $0.07 in the same period in 2016.

Non-GAAP loss per share in the third quarter of 2017 was $0.14 compared to non-GAAP earnings per share of $0.20 reported for the same period in 2016.

Capital expenditures, including purchases of property and equipment as well as capitalized internally used software development costs, were $11.5 million for the third quarter of 2017.

We closed the third quarter of 2017 with $139 million in cash and marketable securities, a decrease of $53.6 million from the quarter ended June 30, 2017. $38.6 million of this decrease was a result of the purchase of nToggle. The balance of the reduction in cash and marketable securities resulted primarily from CapEx investment and from cash used in operations.

Free cash flow for the third quarter of 2017 was negative $15.1 million as compared to positive cash flow of $6.1 million during the second quarter of 2017. This negative cash flow resulted from cash used for operations and capital expenditures. We calculate free cash flow as a net cash provided by operating activities, less capital expenditures, including capitalized software development costs.

The pace of change in our industry experienced this year is unprecedented. As Michael mentioned, we recently decided to be much more aggressive in our pricing and have eliminated our buyer fees, resulting in significant changes to our near-term financial performance.

I will now provide color on some key performance metrics. We believe ad spend will show sequential improvement in Q4 due to seasonality and early signs of positive traction, and estimate it will increase in excess of 20% from the $195 million level reached in Q3.

Please keep in mind there has typical been a seasonal decline from Q4 to Q1. With the removal of all our buyer fees effective November 1, we expect that take rates will decline in Q4 by 400 to 500 basis points from our 16.9% Q3 exit rate.

This move will also likely lower Q1 take rates by an additional 100 to 150 basis points to reflect the full quarter impact.

We continue to estimate that Q4 adjusted EBITDA operating expenses will modestly increase over Q3. CapEx expected for Q4 will be approximately $20 million, consistent with plans and typical CapEx spending patterns.

We are working hard and investing heavily to be the lowest cost and highest value exchange in the industry and are very excited to be in a market-leading position for our customers' needs.

We remain confident and optimistic that our numerous initiatives, including our elimination of our buyer fees, will continue to increase supply, improve our win rate, offer the lowest cost per transaction and drive market share gain.

We are very aware of the short-term financial impact that our recent pricing decreases will have and are carefully monitoring our cash levels.

As we previously discussed, we are for the moment willing to use cash to fund operations and invest in initiatives to drive revenue growth. However, we will be vigilant in managing our costs and are focused on gaining efficiencies through both scale and operations.

I'll now turn the call over to Michael for some closing remarks.

Thanks, David. As discussed, our total take rate for Q3 2017 was 18%. And we exited the quarter slightly lower than 17%. The auction dynamic tests, in combination with the elimination of our buyer fees across our platform, will further reduce take rates in Q4 2017 and into 2018.

In the past, I've noted that we believe the industry take rates will settle between 15% and 20%. However, now we believe that in order to be successful, exchange take rates will be between 10% and 15%.

I am very proud of our accomplishments over the course of the last few months, despite the fact that our near-term results are behind where we hope they would be.

However, with the changes in our industry, we believe that our total addressable market has grown and will continue to grow with further consolidation in ecosystem. So we have made a conscious decision to make bolder moves to capitalize on the opportunity.

We recognize that we are sacrificing near-term growth and profitability, but it's so we can own a larger piece of the more than $500 billion being spend on ads.

The journey to get back to growth will not be an easy one, but we remain confident we have the right strategy and team to deliver.

With that, I thank you for participating.

And operator, you can open the line for Q&A.

(Operator Instructions) And the first questioner today will be Kerry Rice with Needham.

On the take rate, it seems like given your guidance, it's going to settle out 10% to 12% a little bit lower than the -- towards the lower end of your range of what you said the industry is. What -- I guess, what gives you confidence that it settles out there, is the first part of the question. And then the second part is nToggle, does that help at all with the take rate? Or now that you've eliminated the buyer's side of that cost, is nToggle just more of an engagement -- I mean, it certainly helps as a tool for the buy side. But is it more of just an engagement to keep them spending more on the platform versus helping any kind of take rate there? And then I have one quick follow-up question.

Thanks, Kerry. So what gives us confidence why the take rate is going to settle where we think it will. We -- probably interested to know, let David opine as well -- but having been in the marketplace for 15-some-odd years. Now that we have stripped the fee down to just publisher side fees, because that's essentially what's left in how we make money. That business has not been nearly as dynamic as the buy-side fee business; and boy, did we live a little dynamic -- a roller coaster this year with that. That's been very steady. They are our longer-term contracts. The pricing has been very explicit with our selling partners. And if anything, we get the sense through market intelligence that we're actually below-market in our seller fee contracts, giving us some hope that perhaps there's some ability to move and some of our customers are moving it north. But you pointed out, it is a very dynamic market; our certainty is very strong. From a guarantee standpoint, it's always tough to do that. But I really think for the foreseeable future, which would mean 2017 and mid-2018 and beyond, that we've settled where we're going to settle. Of course, there's just one caveat to that, Kerry, and that's mix -- media mix. Some of our products, like orders, carry a lower take rate. And so we'll keep a careful eye on that and give any kind of findings on that, that may have any impact to the take rate. David, do you have anything to add to it?

No, not really.

And then the second question, nToggle. nToggle was always -- we foresee the nToggle as something that can help with win rate a lot more than take rate. And so nothing has changed there. DSPs are getting over burned with traffic -- traffic that they don't even value but they still have to process and incur the costs. So the big hope of nToggle still lives very, very strong. And that is, we can help our buying partners see the -- more of the traffic they want, have to process less of the traffic they don't want. And that will lead to higher win rates on our platform, which, obviously, will then lead to higher spend and net revenue for us.

Great. And then just the follow up is, we heard from one of the buy-side firms, at least for their guidance, as being impacted by the rollout of iOS -- Apple iOS 11. That seems like that would impact the publishers as well, their ability to sell impressions. Have you seen -- and it's early, but have you seen anything there that might impact impression volumes and your ability to sell those impressions?

Yes, terrific question. And obviously, we saw the Criteo results as well. I think that certain buyers are going to be much more impacted upon it; re-mark -- folks that re-target is -- this is going to have a, I think, a disproportion impact to them than say a brand buyer that is looking to target a certain audience. We'll keep a close eye on it. I don't think we've seen anything globally on our platform where we're like, "Oh, that's why that's happening." So to date, we haven't seen the impact of it.

And the next question of today will be Jason Kreyer with Craig-Hallum.

Michael, can you just provide a little bit more color on the decision to eliminate the buyer's fees? Obviously, I'm sure you did some testing, and you ran through some different scenarios. And maybe give us some of the green shoots on your takeaways during that testing that made you more confident, whether it was better win rates, better inventory, any numbers to support that would be great.

Sure. And keep in mind, we eliminated buy-side fees yesterday. So the body of evidence is yet to come. But you correctly alluded to some testing that we did starting in September. That testing was generally around auction dynamics, as we pointed to first -- modified first price, modified second price, which then had a -- the byproduct of that was in instances, sample size didn't include buy-side fees. And when we saw a modified first coupled with no buy-side fee, we saw increase in revenue for publishers, we saw increase in win rate in downstream auctions. And it really emboldened us to say what if we expanded the sample set. And that was the primary driver; that we saw a real opportunity here to grow market share and grow revenue. The secondary driver is, there is no question that buy-side fees, in general, won't exist in the not too distant future. And so we figured let's get out ahead of it. Let's use it to our advantage when others are going to -- other platforms are going to struggle as to how to work their pricing around it. Gain market share when others are slow to adopt no buy-side fees. And in general, lean into it because we think it's transparent, and we think it's the right thing to do for the ecosystem.

And then moving on, the conversations that you're having with your customers regarding server-to-server header bidding. How have those progressed over the last several months? And what is -- what's your perception of market adoption or any thoughts on timing there?

Yes, great question. So I think we've been pretty clear to say that Q4 was going to be -- it's traditionally a locked down time frame for publisher. So any major tech decision, any implementation, is shelved until Q1, Q2 time frame because of the importance of the volume and revenue of Q4. But from what we've heard is, it's kind of a 2-step process, right? Jason, it's open source. Like, so we pitched the concept of why open source is correct, transparency, fair, et cetera, and then our version of open source. And then lastly, implementing on the server. And we have some real live results that I'm sure we'll be able to talk about on our next call about some great engaging conversations with sellers. And I think it's going to move pretty fast, as we get out of the Q4 time frame. There's a lot of folks that are deciding now. And then we'll want to start implementation as soon as possible. So we'll keep you posted but so far, so good. It seems like the bets we've made in this area are going to pay off.

Couple of times that you referenced our model. I think you're referring to the Prebid model. And I just want to make sure, are you still working with some other third parties on server-to-server?

Sure. So it -- part of our choice differentiation for our publishing partners is that if they come to us and say, "Hey, all things being equal, we're deciding to go with the Amazon solution." We support that, and we're close with Amazon and we support -- and we're close with Google. We'll support more closely with anyone that's part of the Prebid.org organization. As long as they code their solution to the standards of the organization, we'll support that as well. Obviously, our preference is our clients to adopt our flavor. But we're going to be very facile as it relates to working with others because we think that's what our clients want. And it will ultimately lead to an increase in ad impressions in the platform, which will downstream results and revenue.

Okay. Just trying to sneak one last one in. Just trying to understand the optics of the write-down on nToggle. Is -- help me understand why you take the goodwill write-down on that one?

Yes, sure. So it's not really a write-down on nToggle. So not to get too deep into the accounting-ese, but from -- for any acquisition that we have, goodwill is pooled into one sort of amorphous group of -- or pool of goodwill. And so we assess that goodwill then. We have a single reporting unit or a single entity. And we assess that goodwill just against the market capitalization and valuation that we went through with our outside valuation firm on a macro basis. So from an nToggle perspective, the minute we purchased nToggle, that goodwill was really -- has kind of separated from the results of nToggle. So we -- our outlook for deployment of the nToggle technology, we're still extremely excited about it. We plan to get that rolled out by the end of this quarter here in the U.S. We don't see any change in the prospects of the potential lift in revenue that will provide us. So it's really separate from prospects of nToggle directly.

And the next questioner will be Mark Kelley with Citi.

This is Nick Jones on for Mark Kelley. Can you walk through the take rate pressure in Q3? So I thought the buyer fees were supposed to be cut in November? Can you also talk a little bit about how we should think about spending growth as the DBA partnership becomes a bigger part of the revenue? And then lastly, any -- are you guys seeing anything from Apple's ITP in Safari?

Sure. I'll take the first one or two. Yes, so in the third quarter, as we talked about last quarter, we -- our exit rate previously was 18.9%. And we expected Q3 to come in something below that. And that held true for the first 2 months of the quarter. We talked about some very significant and broad-based testing that we undertook in September. And that testing was broad-based enough that, that drove a somewhat lower take rate in September than the prior 2 months. And so that's why you have this exit rate of 16.9% and, of course, the average rate for the quarter of 18.1%. And so there was a little bit of drop in that last month, primarily driven by that testing. So that drove Q3 take rate activity. And then, of course, that testing is what led to this decision to eliminate our buyer fees. Sorry, and the second question?

I was just wondering about the spend growth as the DBA partnership becomes a bigger part of the revenue?

I'm sorry, Nick, are you referring to the press release that we issued about DBM and the private marketplace connection?

Yes.

Yes. We think that it will be part of the order story. And what we're seeing as a trend is the biggest advertisers that are buying programmatically are as concerned about environment as they are price. And that's definitely pushing more and more dollars to the order business. And this is just an easier way for them to access our private marketplaces. So it definitely will play a role in the shift that we've seen underway for the last several quarters.

Okay. And then lastly, anything you guys are seeing on Safari's ITP?

Yes. So we had originally addressed it a couple of calls back. We have not seen any direct impact in Q4. Obviously, we deal with thousands of buyers. And so, I guess it's a fair statement to say that collectively, there hasn't been any impact across the programmatic market space. We know that certain buyers that have specialties like re-targeters are probably going to be the first hit from it, but we're keeping a watchful eye on it. But right now, nothing to report.

(Operator Instructions) And the next questioner today will be Sameet Sinha with B. Riley FBR.

It's actually Lee Krowl on for Sameet. First off, I was wondering if you guys can maybe talk about as a percentage of revenue what buyers fees are.

Well, if you sort of look at our exit rate of Q3 of 16.9%. And we're attributing 400 to 500 basis points decrease in this quarter and another 100 to 150 in the next quarter, those decreases are primarily resulting from our elimination of buyer fees. So that kind of gives you an order of magnitude on the scale.

Okay. And then with take rate going down to the 10% to 15% range. And I mean -- to that last point, you guys are kind of taking out fees so that it's a very core offering. Are there any other sorts of friction points, if you will, that could potentially be eliminated that would further pressure that take rate?

No, not -- no. We don't -- we're very, very explicit. If you even go to our website post this earnings, we've posted links to exactly how we make our money, exactly where it comes from. And you'll see that there's no other primary source of revenue other than our seller fees. And as I said before, those are, generally speaking, contractual relationships that are longer in term and very explicit in terms of rate and have kind of held steady for a long period of time here at Rubicon.

Yes. This is Michael. And just for those rates, sell-side rates were heavily discounted because we did have buyer fees. And so there is some opportunity there in the future. We're not going to focus on that from an external results perspective too much right now. But there's some potential upside from that. And then on the downside, there are -- there's a mix, as Michael mentioned earlier, that orders, in particular, has a slightly lower take rate. And so as mix changes, that rate could go down a little bit. But order of magnitude, nothing at all compared to the recent changes that we've seen.

Okay. And then I know you guys announced Google in the last couple of weeks. Amazon is also onboarded. Can you just kind of talk about the partnership landscape and were there other opportunities to integrate?

Sure. I think you hit the 2 largest by far. The other folks that we would work with would be other folks in the open source consortium, the Prebid.org. So anyone that adopts that technology, we've -- as a member of the organization and acting lead role for this year, we've committed to working with partners that have created open source. What we're backing away on is proprietary client-side implementations. We don't feel as though it provides the best monetization experience for our publishing partners and our buyers. And it certainly isn't a great consumer experience in terms of the impact of what occurs with all that stuff on the page. So if it's server side and it's open source, if it's server side CBDA, if it's server side and it's Amazon, we'll lean into it. I would imagine, anyone else has a offering, we'll take a look. I think we want to just be really flexible here. And if we think it's the right thing to do, and it lends transparency, and we have a fair fighting shot in the -- in partnering with someone, we'll do it if our client wants to do it.

And there are no further questions. So this will conclude the question-and-answer session. I would now like to turn the conference back over to Nick Kormeluk for any closing remarks.

Great. Thank you all for joining today. We look forward to speaking with you, again, on our fourth quarter earnings call as well as seeing you at a couple of upcoming financial conferences in January, which is the Citi Conference at CES as well as Needham in the second week of January. Thanks, again.

And the conference is now concluded. Thank you all for attending today's presentation. You may now disconnect.