(Sharewatch) Tremor is an Israel-based company that owns a media buying platform, which allows brands/agencies to buy video advertising inventory using a data driven approach. Video ads are created and paid for by advertisers who want to promote their products. They know that video produces the best engagement versus text or image based ads. Until 2020, Tremor was mostly about serving these customer’s video adverts “out-stream;” eg. a BMW video advert could appear each time a user navigated to a motoring article on The Times online edition. Another large chunk of Tremor’s business also came from serving adverts within an App. Tremor’s system allows brands to decide which impressions to buy and how much to pay for them.
“During the pandemic, one industry that investors counted on taking a hit was advertising. Even digital advertising was expected to be impacted, as negatively affected businesses reduced or paused their ad spend.” says Tremor’s CEO, Ofer Drucker. But every cloud has a silver lining and after an initial wobble in Q1, instead of a crash things just went ballistic. First, advertising began to shift online. Second, more people began working from home, which led to an unprecedented surge in the use of streaming services like Netflix and Amazon Prime - opening a whole new market of placing ads “in-stream’ on Connected TV (CTV). So, a customer might be about to start watching a motor show and before the program is streamed to them, Tremor can serve an ad from a motor company. It can be hyper targeted to allow advertisers to get the most bang for their ad dollar; say you reached the precise bit about BMWs, then your local BMW dealership might put a short advert about their cars in front of you. As a competitor, even Mercedes might be interested in this slot. CTV is explosively profitable and Tremor’s CTV sales went from virtually zero to 25% in Q4 and just before Christmas, Tremor said it would “materially exceed expectations.”
Shift from linear TV to CTV
Although I have cable TV installed at home, I have to confess that I am so long in the tooth that most of the time I watch only the five main channels. I watch a scheduled TV program at the time it is broadcast. This is what is referred to as linear TV, so over Christmas it was rather annoying that the Home Alone movie and several seasons of Antiques Roadtrip were the best of this year’s terrible offerings.
But my children don’t have such problems. They tend to only view stuff on-demand on Netflix or Amazon Prime on their internet enabled CTVs and they binge on their own schedules. When they are not watching on a Smart TV with a direct connection to the internet, they use dongles such as Roku or Amazon Fire, or games consoles. According to one report in March, US households have been spending 29% of their time streaming Netflix during the pandemic. Tremor’s technology helps deliver custom video experiences across all these screens, says Drucker, helping advertisers “tell captivating brand stories to create meaningful, personalized moments” with their prospective customers.
With so many people now no longer watching live television, ad revenue on linear television has begun to shrink as brands increase their exposure to CTV instead. Digital hasn’t killed traditional linear television spend altogether but CTV is starting to cannibalize it. It’s all hugely exciting for Tremor.
Tremor’s recent barnstorming update mirrors the performance of the gorilla in the sector - The Trade Desk (NASDAQ: TTD), which has just reported sharp growth in CTV ad spend, a strong bottom line and a better-than-expected outlook for Q3. That company reported adjusted eps of US$1.27 for the quarter, crushing Wall Street’s estimate of 43 cents. Also, its revenue of US$216.1m came out way above analyst estimates of US$180m.
The first time I had heard of The Trade Desk was when a friend’s daughter went to work at their London office many months ago, having left Channel 4. She had been given some share options and I can remember my friend telling me to take a look at the astonishing share price chart. I suggest you do the same. It has grown to an incredible US$45bn market cap! Or take a look at another rival, Magnite (MGNI) with a US$2.6bn market cap.
At the time, I hadn’t realised I could find something remotely similar listed in London and on a fraction of the valuation. Tremor has been hiding its light under a bushel, in part because the business has been formed from an amalgamation of three businesses (Taptica, RhythmOne and Unruly), each of which had also been highly acquisitive. The company has only Finncap covering it as far as I am aware.
Taptica now 5% of sales
The short history for how the AIM group exists now starts with Taptica. Taptica’s original business was, and still is, in mobile advertising and it works on a pay-per-outcome basis (performance-based). So for instance, Taptica might be called upon by a gaming company to acquire customers for it and it will have negotiated a cost-per-user-acquired (CPA).
In a simple example, Taptica might receive US$100 per user acquired for the client but might have to bid-and-buy US$65 of impressions off its own back to achieve that one conversion, giving a cost of sales of US$65 and a 35% gross margin. The better it targets the advertising and the better the campaign performs, the better its own gross margin is. It was whilst doing this that it began to utilise AI and machine learning to leverage proprietary big data for quality media targeting in high volumes. These days Taptica is 5% of group sales.
Tremor video
The seeds for change were sown when Taptica acquired Tremor in 2017 for US$50m. This brought in a video advertising DSP platform, which nowadays represents the other 95% of sales.
As Drucker explained to me at https:www.scsw.co.uk, the acquisition opened up the “brand advertising” market where B2C businesses use video adverts targeted to mobile/internet video consumers. Rather than being performance marketing activity, this is commission based and it is also a different pot of money - that for building customer awareness of their name and products.
As I am learning, a DSP - demand-side platform - is basically a software platform that allows brands/agencies to decide which ad slots they want to buy and how much to pay for them from one or several marketplaces (called SSPs) on which publishers list advertising inventory. Think of the DSP software as being like a software cockpit where the agency/advertiser can set up their campaign, manage their requirements and see how things are performing. The amount Tremor DSP can pay for any one slot will be set out by the advertiser and the software then goes and does the job of placing their ads in contextually relevant places based on complex algorithms. The price they pay is carried out through real-time bidding by the DSP, determined by supply and demand.
Tremor DSP generates a fee each time a video ad is shown, taking a cut on every dollar that is spent on its platform. The take rate has been steady at around 15-20%.
As a notional example, Nestlé (a Tremor real life customer) might use Tremor DSP when launching a new chocolate. It might have a promotional budget of US$1m, which it wants to spend during July. After liaising with Nestlé, Tremor may arrange to place some of the chocolate video ads “in-stream” with the videos appearing throughout the digital content (eg. you may be watching the news and get served a video advertisement pre-roll, mid-roll and post-roll); or the video could be served “out-stream” in a non-video environment (eg. the user could be reading an article on the Times or Kids National Geographic website and a video will be served by the website’s servers on behalf of Tremor’s client); or the advert could even be served within an App.
To achieve such ad placements, Tremor DSP will look at a number of marketplaces where it can see advertisers’ inventory. When it sees a suitable slot, it has to decide to buy it in literally a fraction of a second between a user selecting to view a video and the video loading, based on the aims of their client’s campaigns.
The situation is complicated by the fact that there could be a lot of other DSPs potentially bidding in real time on the same advertising slot. Tremor’s DSP will use cutting-edge algorithms to assess huge volumes of data in milliseconds to know what is the right price to bid for the slot on these channels. The storehouse of knowledge used to determine the maximum price to pay includes the digital trails we leave on the web, what might have worked in the past and also “first party data” (so in my earlier example, BMW could hook it to its own big data systems). The technology is lightning fast; Drucker says it can handle 3.5-4bn requests to media every hour. In doing so it also remembers results and Tremor uses machine learning algorithms to refine future actions.
When I spoke to Drucker, one of my first questions was how scalable this all is. In 10% of cases, he says that the platform is used on a “self serve” basis, so an agency/brand might enter their requirements onto the Tremor DSP cockpit directly and it does everything for them. But in the majority of cases the placement is managed by one of its 120 account handlers who will provide consultation services to enhance and optimise the video campaigns. So if it looks like the Nestlé campaign isn’t going to be able to place the required adverts during the month of July, they can make tweaks, for instance by raising the bid amounts so that it wins more auctions and places out more ads.
Of course, advertisers also want “brand safety,” to make sure their video does not appear on unsavory sites / context and Tremor DSP will ensure this by using a third party brand safety technology (eg. IES).
Acquisitions give it scale
In 2019, the nature and scale of the video DSP operation was transformed when Taptica underwent an all share merger with AIM-listed RhythmOne.
RhythmOne had its own video DSP called YuMe. Tremor then underwent a process of assimilating the best bits of the YuMe DSP into its own platform, removing costs from Amazon cloud, losing one technology team and removing duplicated costs and surplus offices.
Drucker has since repeated the trick when it bought a third DSP called Unruly last January from News International in an all share deal that resulted in News International owning a 6.5% stake. This time it really hit the jackpot; Unruly had an enviable customer list of big brands including P&G, Nestlé and Unilever and also offices in Japan, UK, Singapore and Germany.
The rationale for both deals is that the digital advertising space is dominated by the “walled gardens” publishers like Facebook. Tremor DSP doesn’t work with Facebook but allows advertisers to target on the “rest of the internet.” The deal with News Corp gave it exclusive access to sell outstream video adverts for 50 News Corp titles including the Sun, the Times and the Wall Street Journal.
It all created a bit of a “halo’ effect. By being associated with those big sites and brands, more publishers now want to work with Tremor, which in turn is helping attract more big name advertisers. A ‘virtuous circle.’ This has been compounded with initiatives such as Tremor’s launch of Private Marketplace Packages (PMPs), providing advertisers with premium supply inventory through owned, controlled and extended packages (15% of H2 20 sales).
If all that isn’t explosive enough, what has further electrified prospects is that Tremor has taken US$23m out of its cost base during the pandemic, making the business hugely profitable.
RhythmOne also added SSL - “full stack”
The story however gets better still. RhythmOne had also brought a platform called RhythmMax into the fold. This is a Supply Side Platform (SSP), which allows websites and video owners to sell advertising space. RhythmOne has agreements in place with 3,000 publishers who connect directly to it. By pushing Tremor Video’s demand-side platform through the supply-side RhythmMax, not only does Tremor get 15-20% of the take from the DSP but it also gets a further 15-20% from the SSP - what Drucker refers to as a “full stack.”
Blue Sky ahead
It’s early days for CTV but Tremor is already profitable and ended 2020 with net cash of US$85m or 64p a share. For the year ended in December, Finncap forecasts sales climbing to US$394m and EBITDA to US$51m (H1 vs H2 EBITDA split of US$1.8m v US$49.1m), for adjusted eps of 14.7 cents. For the current year its forecast is US$470m sales, US$62m ebitda and eps of 21.4 cents. The broker notes that “based on current forecasts, Tremor is trading on 12-month forward EV/EBITDA of 9x on probably very conservative FY21 estimates, versus Magnite and The Trade Desk on 29x and 133x, respectively.” As the Yanks say, you do the math. Buckle up and hold on tight; I am a buyer.
exciting opportunities I’ve seen for a while. I think the price has potential to go through the roof.
