Member Info for 1GW_

There's also a mix input when looking at average CPMs (e.g. static impressions vs Orders vs RTB) which may make it hard to read across from one company to another. Rubicon Project say they shifted from lower CPM static impressions to higher CPM Orders and RTB. They also associate higher CPMs with their recently introduced "Fastlane" header bidding feature. Rocket Fuel also refers to header bidding in its call - seems its an industry talking point - hope RhythmOne has a good product/solution. This from the Seeking Alpha transcript of the Rubicon Project call: "At the same time, we continue to further innovate in our core business and we see solid increased demand for our RTB products, which remains a key growth driver for 2016. Our recently introduced FastLane header bidding feature is particularly relevant to both our RTB open auction and Orders offerings. FastLane enables Rubicon Project to sit much higher in the ad stack of the web publisher or mobile app, which allows the strength and scale of our buyer demand to create much greater demand for impressions. This leads to higher CPM prices and increased in revenue for our seller clients and greater impression volumes to Rubicon Project. Early results from FastLane adoption are promising with many of our largest sellers either already on-boarding or in the queue for Q1. Sellers such as Tribune Publishing, a long time partner, which operates prominent outlets such as the Chicago Tribune and The Log Angeles Times has been using our Orders platform exclusively. So it was only natural that Tribune turn to our FastLane innovation to drive a more holistic approach to managing yield on their properties across all types of inventory, including the most premium and highly desirable impressions for which buyers are willing to pay more. Since deploying FastLane, Tribune Publishing has experienced a significant increase in their CPMs while delivering better ROI for its advertisers." hTtp://seekingalpha.com/article/3923026-rubicon-project-rubi-frank-addante-q4-2015-results-earnings-call-transcript?part=single

Big step up in ComScore uniques and reach in January. Apr 158m 68.6% (month, uniques, reach) May 159m 68.8% Jun 157m 67.6% Jul 157m 67.7% Aug 148m 63.4% Sep 145m 62.4% Oct 149m 64.2% Nov 152m 65.3% Dec 161m 69.5% Jan 178m 76.2%

A couple of other thoughts on this ahead of the ComScore data: ComScore is desktop only, and Quantcast doesn't seem to be able to identify those people who access sites only on mobile (it can split online from mobile web unique devices, but doesn't do this for people). So I would expect Quantcast people to be directionally higher than ComScore desktop unique visitors because of those Quantcast people who only ever use mobile web. What would be really nice is to see R1 start to appear in the ComScore list of top DSP/SSP/Ad Exchange entities. Not quite sure how these are determined (there are only 3 in the December list) but it sounds like the sort of list R1 should be aiming for.

I would be surprised if the ComScore January number (unique visitors) gets anywhere near as high as the current Quantcast people number. For September 2014, Quantcast was reporting about 200m monthly people for Burst Media network, when ComScore had unique visitors at 166m. More recently, google.com is reported as 232m in Quantcast (albeit not a quantified site) for December 2015, but 200m in ComScore. So clearly there is a methodology gap if they're trying to measure the same thing. And as has been said many times I think, the current Quantcast numbers do not reflect a sudden increase in visitors to R1 sites, they just reflect the fact that these visitors are now being counted in the public Quantcast stats - and in Quantcast the RMax sites only started being counted towards the end of January so a lot of the increase in reported stats came through in February, not January. Having said that, Quantcast appears to indicate an increase of about 60m people since the AGM (150m reported in AGM slide based on non-public data, 213m people being reported now in the public data). So assuming this is a genuine increase in reach, we might hope most of it would show up in ComScore at some point. December ComScore at 161m uniques was up just 16m on the September low of 145m. Lies, damned lies and statistics, remember...

The blog seems to confirm that R1 is now using Quantcast to measure total reach, rather than using it just to measure current campaigns. By its acknowledgement of Quantcast as a ranking tool it suggests to me that it believes its peers are using it in the same way. Interesting to see that they chose "classic view" in the graphic.

I've just bought some at 25.00. I have a relatively big holding in Crest Nicholson and a smaller holding in Telford Homes and have been watching Bellway for a while, attracted by the low p/e and the growth profile. It seems (to me) due for a bit of a bounce and I'm hoping tomorrow's TU can provide the catalyst, although Redrow's share price performance today following their interims isn't very encouraging so far. I guess one question mark over Bellway might be their exposure to "high-priced London flats" through their Nine Elms position, so hoping they can find a way to indicate something positive tomorrow about how they see the London market.

The October 2014 data in the earlier messages comes from an "Inside Quantcast" article. In the "Explore" box, search for "Top Global" and you should get a link to the article which contains a list of the top 20 global ad networks and platforms in October 2014. What is thought-provoking is the change between the data in the table and the data shown for the same period in the current Quantcast stats. If I look at the top 3 networks in the October list for which data is still available (i.e. Rubicon Project, OpenX and sovrn), the data is almost the same (e.g. 168m US people in the table for Rubicon Project, compared to the current Quantcast data now reporting 170m for Oct'14). This suggests to me there haven't been any major methodology changes which caused major data restates. But if I look at Burst Media it was at 147m people in the list, whereas now R1 is reported as just 114m people for that month. If I look at data for other months I recorded at the time vs the data in the database now, there are similar or even bigger discrepancies: Sep 14: 202m at the time vs 121m (for R1) now Aug 14: 159m at the time vs 105m now Jul 14: 89m at the time vs 62m now So the data reported at the time for Burst Media network was significantly higher than the data now reported for the same period for RhythmOne Network. The question is why? Could it be filtering? Historic results reduced to reflect the proportion of traffic currently filtered out. Could it be disposals? Historic results reduced to reflect parts of Burst Media network no longer in RhythmOne or marked for future disposal. Whatever the reason, it suggests to me that the current stats may be more impressive than I first thought, because the historic comparators now appear significantly lower. Nothing to do with monetisation or revenue of course, but struck me as worthy of comment.

7th in the US and 8th globally! However, that was for the official October 2014 full-month profile. Burstmedia network reached a peak I think of 204m US monthly people in the data shown on 26th September 2014 the 30 days to 25th September I guess). At this point it was ahead of google.com which was on 203.7m monthly people. It had overtaken OpenX (number 4 US rank in the October listing) and possibly sovrn (number 3), but I am not sure it ever caught Adblade or SpotXchange. On current trends it now looks reasonably likely to me that R1 will surpass this 204m peak.

SP180 - I tend to agree that a loss in translation is the most likely explanation. If the RMax site stat really is quantcast "uniques" (devices) it seems suspicious that the stat is also 150m in the other 2 cases, one of which is unambiguously "people" (IMO) and even the RMax press release talks about "unique users". On Burst Media stats, they got as high as 75% reach and over 170m unique visitors (i.e. "people") on the ComScore stats in November 2013. They were so proud of this that they talked on the Burst Media site about reaching three out of every four online unique visitors in the US. The RhythmOne Quantcast monthly stats should keep climbing for the next 3 weeks or so as we build up to a full month's data from RMax - so the US monthly people stat (currently 166m) should be considerably higher (and much higher than the 150m quoted at the AGM) before the month is finished, which will presumably reflect in large part the third party supply added through RMax since August as you say. Monthly people are currently increasing at about 5m/day (over the last few days) but I expect there will be diminishing returns (i.e. fewer new people and more repeat people) as we go through the month.

On Quantcast, blinkx splits the data between RhythmXchange (RMax) and RhythmOne Premium. RhythmOne Premium (look at the subsegments) seems to look in the main like the old Burst Media vertical channels of Giant Realm, Stadium, MomIQ, Ella. So I just wonder if one of the reasons for confusion over the various 150m stats is that blinkx is counting these two segments separately. So the 150m for the 1R platform in the AGM slide could have been combined reach (monthly people) of both segments, whereas the "150m uniques" and "150m unique users" on the RMax site and in the RMax press release could have been uniques (rather than people) just for the RMax segment. Does that make sense or not? A curiosity of the RhythmOne Premium stats is the very low mobile % (only about 3% of daily uniques on average so far I think). Yet I have an old flier for Giant Realm which talked about a display audience of 49.9m monthly uniques and a Mobile Audience of 53.2m monthly uniques (both stats from ComScore in Sep 2014). So at the time Giant Realm at least had a Mobile Audience that was bigger than its Display audience. Does that mean that Quantcast is not yet picking up the mobile audience properly for this segment, or has blinkx taken a lot of the mobile offline while all those long-tail sites are optimised for mobile? Either way it seems like there might be potential for a big increase in the mobile stats associated with RhythmOne Premium. The sooner we get that blog from Dan explaining how to interpret the stats, the better...

Shroder - I thought we'd agreed there was some imprecision in their terminology! "150m uniques" Rmax site vs "150m unique users a month" in RMax press release vs "150m" on AGM slide. The only "150m" which appears unambiguous to me is the AGM one because of the reference to 50% of the US population. However, we're now well past the 150m on both "uniques" and "people" in Quantcast terminology, so it's a step forward whichever way we choose to interpret it. We need 30 days of RMax data on Quantcast to see whereabouts it's going to top out and so far I think we have only 7. So another 23 days of the current sort of run rate ought to get us up to a fairly impressive people number.

I agree the people metric is an estimate and in the past have thought of Quantcast stats as a whole pretty useless as a measure of blinkx' success because of blinkx' ability to turn tags on and off to match their campaigns. BUT Brian Mukherjee cited Quantcast at the AGM as the way to follow RhythmOne's progress and there has been nothing from the company since to contradict it. Shroder has suggested blinkx might offer up some commentary on the stats (" further discussion around measurement stats via their blog posts in due course.") and I am looking forward to seeing this. In terms of the people stat in particular, I don't know - maybe blinkx will downplay it in the blog. But they seem to me to have quoted it at the AGM and Quantcast appear to use it to rank so I assume both believe it has some validity.

People "The estimated number of people accessing a property from online and mobile web in aggregate" Uniques "The number of mobile app users and online cookies for a property" "The People number aims to describe the actual number of human beings viewing a web property: One of the several issues with using the pure ‘cookie-counting’ methodology is that it does not account for the same person visiting the same site from multiple computers. The same person is visiting the site, but they are being counted as if they were two." All quotes from Quantcast site. The interest in Quantcast is that Brian Mukherjee was asked at the AGM how we can best follow the progress of RhythmOne from the outside and he said Quantcast. The 150m stat quoted in the AGM slide is I believe a "People" stat (because of the reference to 50% of the US population) and the "People" stat is what Quantcast appears to use to rank different sites.

Well that's a relief - I was beginning to think they'd end up stuck with a loss-making Kitbag. Bizarre timing though given they put the TU out only on Friday - unless there was some regulatory issue about getting it out in January? Simplifies the game a lot I think now. Sports Direct now just has to decide whether it can live with its hurt pride (bluff called on getting rid of Kitbag) and if so is it sufficiently interested in the rest of Findel to go for a full bid. If not, then could other companies be tempted now that the threat of a messy Kitbag-related battle with Sports Direct has gone.

152m in fact. So it has now overtaken the AGM stat from August and looks like it should continue rising for a while.

For "people" you have to use the "classic" view - click on the "measure classic" box.

And at 146m monthly people, almost up to the 150m quoted for August on the AGM slide! Still, if the current trend continues we should overtake that very soon.

I suspect the "150m uniques" quoted on the RhythmMax site and the "150m unique users a month" quoted in the RhythmMax launch press release are the same as the "150m" for the 1R platform quoted in the AGM slide. This last 150m is talking about quantcast "people" not quantcast "uniques" (i.e. unique devices) I think because it refers to 50% of the US population. So I think the quantcast public data (130m people) is still showing us below this level. However, there's only 2 days of data for the RhythmXchange (RhythmMax) segment, 2m people on the 27th Jan and 11m people on the 28th. It looks to me like "Burst Media" tags were removed on the 20th/21st Jan and RhythmMax tags activated on 27th Jan, RhythmOne Premium on the 20th. So we're still transitioning from one to the other and it's a bit early to extrapolate to a full month's data. But I agree, the trend is encouraging at this early stage!

Steady on. The key metric is monthly people rather than monthly uniques I believe. At the AGM blinkx showed a slide of August 2015 quantcast data where the "1R Platform" was 150m and the slide was tagged with "The 1R platform already reaches c50% of the US population". US population seems to be about 320m from a quick search so 150m being about 50% seems about right. Now, Rhythmone is currently shown as 130m people on the public quantcast site, so 20m fewer than on the AGM slide. So I think what's happening is more of the internal blinkx tagging of sites is being made public and this may be the main factor driving the impressive recent increase. For context, in September 2014 Burst Media reached around 200m people. For further context, some of the higher placed ad networks that I have found by trial and error are: Pulsepoint (240m people, 986m uniques, rank 2) SpotX Video (237m people, 2,695m uniques, rank 6) OpenX (237m people, 720m uniques, rank 7) SOVRN (217m people, 527m uniques, rank 10) At the moment AdKarma network (34m people, 41m uniques, rank 72) still seems to be shown independently of RhythmOne, so that might give the RhythmOne stats a further boost if it is integrated. So while it's good to see the quantcast public stats get to the sort of level that blinkx appear to have been seeing privately in August, I'm not sure we can infer that visits to RhythmOne sites are increasing at any particular rate. It would be nice to think that at some point blinkx might let it be known that quantcast public data does reflect the actual reach of the RhythmOne platform, and at that point it would make more sense to take notice of the trends in the quantcast data.

There's a turnup. Mid-morning profit-warning but the share price manages to get itself into the green towards the end of the day. Intriguing.