This study was conducted in 2009. Since then, we expect the increase in multi-screening behaviour and internet device penetration will have increased online response to TV ads.


We all know that TV makes people do things. From talking about last night's must see programme to hitting the red-button and voting for your favourite reality contestant or even joining in with Gordon's Cook-along, most of us regularly respond to the content on our screens in one way or another.

The advancement of technology has only served to further our ability to respond to television. We can now interact through the red/green buttons and increasingly through the phone (particularly mobile) and, more commonly, the internet. Advertisers have always been keen to harness viewers' willingness to act upon TV content and driving them online is often now a key objective. Our research 'TV and Online: better together' conducted with the IAB, demonstrated that advertising can drive people online in a host of ways, the benefit being that the consumer journey is now much shorter and viewers can now respond to TV ads as and when they happen. TV really has become a point-of-sale medium.

However, the shifting technological landscape means the traditional 'rules' of response planning are becoming blurred since the days there were first created in the landmark study by Channel 4 and BT in the 1990s. Thinkbox therefore sought to 'rewrite' these rules by exploring the role of TV in generating short-term response, particularly online, and ultimately, defining TV's impact on short-term return on investment.

We already know that TV pays back in the long term

There have been many studies in recent years that detail TV's effectiveness in the long-term. Our own econometric 'payback study' with Pricewaterhouse Coopers demonstrated that on average, TV yielded a £4.5m sales increase for every £1m invested - significantly more than any other media.

Figure 1: TV Delivers the best long-term ROI.

TV also affected sales in the year after investment almost as much as during the year in which the activity was broadcast, which can pose problems when it comes to measurement as all too often those effects are attributed to whatever media are active at the time. TV investment was also the main driver of large brand values across a wide range of different categories. The IPA's 'Marketing in the Era of Accountability' econometric analysis on 880 IPA Effectiveness case-studies also supported these results.

But when it comes to TV's ability to drive short-term value, the waters muddy. As broadcast media have no common return path, TV is all too often seen as being less accountable in spite of the barrage of evidence to the contrary. The fact is, TV is no less accountable, TV is just harder to count! Consequently, when people have talked about 'TV's declining effectiveness', they usually mean that lower response rates via the dedicated telephone channel to their DRTV campaigns.

What about TV's role in driving response?

Research by Mediacom demonstrated this point. Superficial response analysis showed that DRTV was amongst one of the most expensive media channels in the media mix in terms of the cost per customer acquisition and that banners and search were the cheapest. However, after reducing DR investment in TV and shifting it online, the picture reversed. It became apparent that the offline channels, particularly television, were driving consumers online but that the online part of the customer journey was being credited with driving the final response. Mediacom therefore applied their detailed econometric analysis. The relative cost of DRTV fell dramatically, with TV often becoming as cheap as - and in some cases, cheaper - than online in terms of the true cost of acquiring each customer.

In addition many DR advertisers have witnessed significant declines in phone response over the past 5 years (which is unsurprising considering the increase in online access over the past decade) but again this is rarely captured in traditional response analysis. Analytic measures that tease out the true value of TV in relation to short-term response therefore need to be applied.

Understanding online response to TV advertising

We commissioned Mediacom to examine this further and merge the findings with what we already know about TV and response to give us a more holistic view of TV performance. We also needed to measure the immediate online effect of TV advertising to enable practitioners to better optimise their campaigns at a spot level.

Mediacom analysis revealed that TV typically accounts for a third of ad-driven sales, although TV spend usually accounts for less than this. The proportions vary by sector, brand and levels of investment.

Figure 2: We know that TV drives a third of ad-driven sales

Advertising also creates a long-term 'base' of both sales and response that varies between 50% and 300% of the short-term effect - again dependant on brand and sector. But what about the immediate effect of TV advertising on web activity?

The methodology and top-line findings

Mediacom devised an innovative econometric technique to link TV spots to the initial web activity conducted within a ten minute timeframe of the broadcast (it should be noted that influential factors such as typical web usage, exposure to other media etc, were factored out). In some cases, deeper levels of web interaction were measured, such as downloading a brochure, filling in a form or requesting more information.

Seven brands were recruited across six different categories including finance (two brands), automotive, travel, consumer electronics, charity and telecoms. The categories and brands were selected to be representative of the DR market, but also to represent the different strengths of the calls-to-action featured on TV ads. All in all, over 175,000 TV spots were included within the analysis.

The research revealed that for the relevant participating brands, 46% of the total response was campaign driven. The remaining 54% was the underlying base, driven by factors such as long-term advertising effects, brand equity and consumer habits.

However, of the 46% of response that was ad-driven, just under half of that was driven by TV advertising - marginally less than the effect of all other media combined.

Figure 3: TV drives half of all campaign-driven response

In addition, across the duration of a campaign, 60% of measurable response (so response through channels such as online or unique phone numbers) came via the web, rather than the phone - with a third of that happening immediately. Mediacom had already identified that just 15% of web response happened immediately over the phone, which proved that practitioners are currently optimising based on just a small fraction of the total measurable response. We were therefore able to deconstruct levels of response against all of the different planning elements to help 'rewrite' the existing rules.

Figure 4: How big is immediate web response?

Rewriting the rules of response

Back in the late 1990s, Channel 4, along with BT, commissioned a piece of research to measure the number of ads that carried telephone numbers and assimilate this back to TV spots and subsequent levels of response to those TV ads. The work was pioneering and has provided the framework for direct response TV planning to this day. However, the web is now the most common response vehicle and there is a real need to rewrite the rules to fully develop TV's potential to drive viewers online.

Rule 1: weekends are just as responsive as weekdays

Conventional planning wisdom relating to phone response shows weekdays to be much more effective, with response tailing off as the week progresses. (This is partially due to call-centre availability at weekends, but is also a symptom of the mindset many people are in, where the idea of holding to a call-centre seems especially unattractive in their free time.)

However, web activity shows a different picture to phone, with weekends particularly responsive and actually peaking on Sunday - the worst day of the week for generating phone response. It would therefore be beneficial to advertisers to consider tailoring their creative and running web-specific copy at the weekends

Figure 5: web response is just as high at weekends as it is during the week

Rule 2: DR campaigns can run into later dayparts, particularly for high interest categories

Phone analysis shows response falls as the day progresses and that earlier dayparts are key for driving consumer activity.

However, the reverse is true for web - response actually builds throughout the day, climaxing in the peak dayparts. (It should be noted that breakfast time is not ineffective, it's just less likely to drive instantaneous response as viewers tend to be preparing for work or commuting). Therefore, if the additional cost of the airtime permits it, evenings could be fertile ground for DRTV.

Figure 6: evenings are the most responsive for web

Web response is also much flatter than the phone throughout the weekend, particularly on Sunday, where it holds up much better in peak.

Interestingly, there is a real difference in web activity between the high and low interest categories. For low interest brands (such as insurance), web response is much likely to occur earlier in the day, however the picture is reversed for high interest categories such as cars and travel. Again this reflects the mindset of viewers at different times of the day. In the evening, we're much less inclined to perform tasks we view as chores, whereas those we enjoy are a different matter. In addition, TV is generally a shared experience and provides an ideal environment within which to discuss shared decisions with the rest of the household.

Figure 7: Response in low interest categories peaks earlier in the day

Rule 3: Centre breaks and end-breaks can both be effective, depending on the level of viewer interaction

The work on commercial break ecology by Billetts in the 1990s revealed that centre breaks were better in terms of recall. This was mirrored in terms of web responses to homepages, with a higher index of uplift per rating. However, the reverse was true when it came to further web interactions such as entering personal details or downloading information, where end breaks became the most responsive. We can therefore conclude that it's highly likely that initial interactions take place during the midst of a programme, but a viewer will wait until their programme has finished before interacting more fully with a site.

Figure 8: centre breaks and end breaks serve different roles depending on the level of viewer interaction

Rule 4: 1st, 2nd and last in break are the most responsive.

Again, based on the Billetts work, first, second and last in break are deemed premium positions due to their ability to generate higher levels of recall. Levels of web response per rating confirmed this, suggesting that the benefits of increased response for these positions outweigh the additional monetary investment needed to secure them.

Rule 5: longer time-lengths can work well for DR campaigns

Generally, campaign time-lengths are dictated by communication objective. For ads that have a strong story or complicated propositions, ads tend to be over 30 seconds long, but for simple, 'reminder' messages, shorter executions are more common. On the whole, as with brand ads, most response ads are 30" long.

However, one advertiser within this research (Audi) utilised a 60" creative execution. Although response was not a primary objective for them (for example, there was no call to action anywhere in the ad), they had noticed sharp spikes in their activity when their brand activity was on air and wanted to investigate this further.

Interestingly, the 60" execution proved far more responsive than the 30" cut-down of the same ad, even when the relative cost of the airtime was factored in. This highlights the additional engagement that can be harvested through longer time-lengths for strong creative executions.

Figure 9: the longer time-length roved more responsive

We can therefore conclude that, although not statistically robust, longer time-lengths can work well for response in the right circumstances, eg. a great ad or a complex message, although advertisers should still use the shortest time-length to get their message across effectively.

Rule 6: Creative Copy

Levels of response can differ by creative copy. At one end of the spectrum, there are brand ads that have no call to action whatsoever and are usually aimed at making viewers feel, rather than do, something about the brand. At the other end are pure DRTV ads which have a strong and focussed call to action and are aiming to drive consumer behaviour in a very explicit way, for example, calling a unique telephone number or logging on to specific domain. Somewhere in the middle, there is brand response, which is about driving brand perceptions but usually with a discreet call-to-action such as a web address.

Conversely, Mediacom tend to see a higher response uplift per rating for their brand activity via the phone, but this is down to the higher quality of airtime bought for brand response (BR) campaigns over direct response. Unsurprisingly, when airtime costs are factored in, DR copies become more efficient.

This pattern was replicated for the web, but another phenomenon was emerged. Brands ads, that had absolutely no call to action, were shown to drive instant response online - something that was never a possibility before web became a key response vehicle.

Figure 10: even pure brands ads on TV can generate a response online

There are several successful instances of high profile advertisers using brand ads as the spine of their campaigns and switching response on later (such as the T Mobile 'Life's for Sharing' ads). However, as seen from the above chart, brand ads were six times more effective at driving immediate web response online than pure brand executions. We would therefore urge all brand advertisers to consider adding a discreet call to action to maximise the potential flow of response.

What about genre?

Although we examined channel genre, the results were non-conclusive. Results were completely dependant on the type of brand and its category, the strength of the creative, the style of the message and the campaign strategy.

In summary

From the analysis it has become apparent that the role of TV in driving response, particularly online, has been massively undervalued. Partially, this is down to the changing technological landscape. As phone rates decline, the web becomes more dominant but until now the effects have not been captured. Indeed TV's impact on online - and indeed the impact of offline media overall - all too often slip under the radar because of the more advanced measurement techniques needed in order to understand the bigger picture.

The reality is this; all TV is now response TV. All ads shown on the broadcast stream can now be instantly searched, shared, copied and appraised at the click of a mouse; transcending the past boundaries of TV and adding a whole new dimension to response advertising.

TV Response: the new rules

This econometric study from MediaCom, commissioned by Thinkbox, measured the impact of television on both immediate web response and other short-term response channels. It’s the first time that the instant effect TV ads have on web response have been measured and made publicly available. Here you can catch up with the background, methodology and findings of this research, plus some “new rules of response” which we’ve put together for you. These come to you in the form of some best practice advice, including how to optimise your creative and media activity to capture the response that TV generates. All in all, this study reveals that the role of TV in driving response, particularly online, has been massively undervalued.

Register here

How TV works, why it works, and how it will develop in the future. Register here to keep in touch with the latest from Thinkbox’s rolling research programme, find out about our launch events, and also to access all areas of our site.

Associated Content

The accountancy and consulting firm Deloitte has just published their 2010's media predictions report which focuses on the consequences of technological change - particularly digitization - and are shaped by 2010's economic outlook. Amongst a wide range of topics, they address the demand for on-demand TV, the integration of television and the web and the short-term prospects for 3D television. Here you can get the top line on their findings and also link through to their full report. It's well worth a read.