Tracking Audio Advertising
Our portfolio company Targetspot announced something pretty interesting today. They are calling it Targetspot Analytics and the announcement is here.
When you run an ad campaign on the web, you can track its performance. That's been a vital part of the online advertising value proposition from the early days of web advertising. We can argue about what metrics are worth tracking, and we do, but the ability to measure online advertising is a key reason it is growing and other forms of advertising are in decline.
But it has been hard to measure the "performance" of video and audio advertising, particularly "in stream" advertising like pre-rolls and mid-rolls. That is because users don't click on the ads and leave the stream.
But we know that radio and video advertising works. People see and hear things and they remember them. I listen to WEHM on long island all summer and I know where to get outdoor furniture, get a tennis court resurfaced, and where to get my bike tuned up because I listen to the ads in between the music.
So back to Targetspot Analytics. This is how it works:
TargetSpot Analytics tracks the unique and total visits to any
advertiser-designated destination website along with other data,
following the delivery of an online radio ad. This information gives
agencies and advertisers the ability to measure a true return on
investment and optimize the performance of their online radio campaigns.TargetSpot Analytics is easily implemented: An advertiser simply adds a
small snippet of code to their destination website that corresponds to
their TargetSpot ad campaign. TargetSpot Analytics can then provide
information on each visit to the advertiser’s website, even if the
consumer arrives there later through a search engine or by typing the
website directly into the browser.
The bike store on long island doesn't know that I showed up to get my bike tuned up because of their ad on WEHM but if they ran that ad on WEHM's internet stream and I went to their website to look up their location before heading over, they would. That kind of tracking and measurement is the power of internet advertising in action and I am excited to see it come to audio and video in stream advertising.
![Reblog this post [with Zemanta]](http://img.zemanta.com/reblog_b.png?x-id=8f6978d3-08eb-4165-acce-f1452b01def9)
Comments (Archived):
sounds like an interesting concept. quick question reg. how everything is connected. does the internet radio company (i.e. pandora) place a cookie on my computer when an audio ad plays?
i’ll ask targetspot to respond to this question
Happy to answer your question with the technical detail.We use a similar technique as is used to track and measure display advertising. It involves dropping a cookie and then looking for that cookie to show up on the advertiser’s website.The Internet radio companies will not, in fact, do anything. We do the the entire implementation ourselves. So yes, when an audio ad is delivered we associate a cookie with it. If our advertiser chooses to integrate our code on their own web site, then they will be able to measure which of their visitors are users that have listened to their online audio ad through us. That is extremely valuable since it allows our advertisers to measure the effectiveness of their online radio campaigns (much like they do today in display ads, for example). This, in turn, allows them to allocate their budgets to the best performing audio creative, most successful targeting criteria, etc.We also support any 3rd-party tracking, such as DoubleClik DART or Microsoft ATLAS, which would work in an indentical manner. If the advertiser prefers to use its own 3rd party tracking tool, we will simply drop for them their DART or ATLAS cookie, and they can track performance through those 3rd-party analytic systems. While our new proprietary analytics tool provides the convenience of one-stop shop to audio advertisers, our philosophy is about prooviding advertisers choice and flexibility to work with the analytics platform that most suits their needs.Eyal GoldwergerCEO, TargetSpot
So the cookie is dropped on the user at the time that the tracked ad is played by the broadcaster? Very interesting: is your code tracking stream metadata to know when the ad happens, or are you taking a different approach?
The code inserted on the “destination” website of an advertiser can read and attribute back to our servers proof of activity from the cookie. That’s the crux of it. Where we seat in the stream allows us to do this, and that is the novelty.
i’ll ask targetspot to respond to this question
I thought you listened to NPR and avoided commercial radio. But the technology this company is implementing is valuable.In the print/TV media measurement is all gray and I guess this company may remove a lot of the grayness in measurements and provide value to the advertisers.
Not just classical advertising – could also be very interesting if applied to public information type radio services to see what resonates with the listener.
If you use doubleclick/Atlas you can currently do this with spotlight or atlas tags – how is this different than those two options?
It is essentially the same. We give you the option to choose between our TargetSpot Analytics product, or ant other 3rd party tracking product, with a philosophy of providing choice and flexiblity to the advertiser (the former option provides one-stop shop, the latter provides support of current workflow – whichever you prefer). The novely is the support of such capability, common in other media, in online audio advertising – here, this is a dramatic shift of the conversation and a new way for advertisers to be spending budgets and tracking their campaigns.
I believe the most effective audio advertising is when listening to talk radio. Sports and Howard Stern spots are big money makers and… effective. I actually pay attention to ads for these segments but for music… not so much.
agreed. the absolute best kind of talk radio advertising IMHO is when you can get the trusted broadcaster (i.e. howard stern) to say the ad for you.
Does this really show the ‘true return on investment’ for the advertiser? What if you never go to the bike shop’s website – but rather directly to the physical shop? What if after hearing the ad on your laptop, you use your uncookied iPhone to navigate to the site’s shop, or to respond to a friend who asks if you know a bike store on LI? What about word of mouth? Don’t all of these also have real value to the advertiser?Clearly online advertising has brought powerful tools to allow advertisers to measure ROI in ways that traditional media couldn’t, and that’s generally positive, but I’m concerned that if the online advertising industry goes too far in this we-can-measure-all-ROI-via-conversion-metrics direction – removing any acknowledgment of the value of getting your brand in front of a valuable audience, even if they don’t click through – that both distorts reality and cheapens the value that online publishers bring by virtue of their audience. It’s online advertising as direct response junk mail – a point Jim Spanfeller made very well here:http://paidcontent.org/arti…
yes. all you mentioned is even more important when we consider how so much of social media is essentially about word of mouth. the best ads are the ones people talk about.there is also the huge issue of integrity of these analytics. in addition to all the stuff it is admittedly not counting, what about the stuff it says it is counting but sometimes gets wrong? the margin of error on analytics can be quite large in my opinion.
This is a fantastic innovation. Just one more reason for advertisers to pay attention to in-stream audio advertising. Congrats Fred and Eyal.
My Question of Choice: So I listen to Pandora, usually Late Romantic Composers, instrumental only, in a Dormitory Lab at College. I don’t usually play just for me: It’s usually for me and my friends ( it’s a good mix for typing up papers in a large group, among other study behaviors)Since I suspect your analytics would probably tell you about my computer (named Ezekiel, Ezzie for short), and not anyone else who is listening (and that can be a larger amount of people), how would you compensate to be more effective. Especially because this is something driven by metrics. So far we only have me, not anyone else passing by. Does your cookie track say my link passage of what music from these stations, or similar, that I am doing at the same time, in order to calibrate for “pretend audiences” that may or may not exist?(FYI, I’m not the only person who does this. Group listening to music is a national pastime. My favorite campus coffee shop has each barista blend together Music stations for the coffee shop using Pandora as well…I kept trying to figure out where there music was coming from…and they won’t share the stations either…)
Good point/s.I believe the advertising industry is still in its infancy regarding understanding (if it ever will) the web, let alone streamed music services, etc. Even TV became a confusing medium for it with the advent of VHS, TiVo, etc. I don’t think it’s ever recovered from that. Who knows, there may be a renaissance in print-advertising one of these days ….It still operates on a scatter-gun principle and vastly over-rates itself in its significance/leverage.I only listen to BBC radio stations and virtually never watch real-time TV, so the whole premise is somewhat irrelevant to me anyway! Oh, and in well over a decade of active web use I have never clicked on an advert …. ;-)PS, the floundering of web/advertising is summed up by Facebook inviting me to see ‘More Ads’ – yeah, sure, great, thanks! – and now also asking me if I ‘Like’ an advert(!).Unbelievable.
I just figured I was doing the typical thing. and that typical thing is metrically problematic, for you’re trying to measure whose walking by one computer…
Metrics are challenged with getting an average return and an error bound for a given expenditure. I love a good detection problem, and the solution is to attack the problem from several directions.While tracking cookies may seem like a good measure, it’s the equivalent of asking an audience to get a barcode on the back of it’s neck. Most folks are resistant to adding extra tracking software, and it’s really limited to a single connection.A simple single question survey at the bottom of a landing page will get interested users to give feedback. If that doesn’t work incentivize customers with discounts to identify the referring source. I’d gladly click on a multiple choice tab or tell a cashier a referring source for a 1% disount.A subset of users who already are comfortable sharing their experience with businesses and ads is having a great time sharing their experiences online 24/7. We need to mine the stream with advanced processing algorithms for signs of effective ads.
Actually I’m wondering if you know where to get a bike repaired in our neighbourhood the one near me closed.
Can we put earplugs on every human being and track what sounds are getting in…:)