Commentary
Data-Driven Creative
- by Michael Kaushansky , January 17, 2012
As we start the new year, we are entering the age of data-driven creative strategies. Including data as an input into the creative cycle will allow current trends, consumer sentiment,
real-time behavior and attitudinal response to be interwoven into creative thinking, thus making a campaign more relevant, dynamic and effective.
Data can enhance creative in major ways throughout the creative lifecycle. Here are three major ways data can help:
1. Help with design. The science of understanding how brands and products
are perceived by the consumer is as old as marketing itself; however, current techniques allow marketers to quickly get a real-time pulse and dive deeper into the consumers’ psyche. The
following tools can provide help with design:
Syndicated research – data available through a host of
syndicated tools, including comScore, MRI and Simmons will start to profile demographics and behaviors of your target audience.
Social listening – unbiased scrape of the social web to get a pulse on buzz, sentiment, and sharability of a trend, brand or product. Get a sense of what is being discussed, what might the completion be doing and what is perceived to be old news. Are influencers promoting these discussions?
Search insights – pick-up the latest searched trends and keywords. Evaluate the trends by segment (region, time-frame, industry, audience or category). Look to see if your creative buzz words are already being searched for?
Website deep dive – if you focus research around a few coveted websites, see if traffic is picking-up/down. Who is coming to these Web properties, and from where?
Panels & polls – selected panels that operate as sandboxes for bouncing new ideas. Polls are more efficient than focus groups since you can poll hundreds -- even thousands -- of users on key sites (e.g. Facebook, LinkedIn, etc.) quickly and virtually for free
2. Help with pre-market concepts. Once developed, creative concepts need to be aligned with media and the individual consumer -- in other words, personalized. Personalized creative can be as simple as aligning creative concepts to specific media channels, since we know media consumption varies from person to person. A more advanced option is to personalize creative to the consumer; this has been done before. Check out your latest credit card offer from Chase. I bet it has your name and a lot more detail than a display ad has on your favorite site.
Among the tools helpful at this stage:
Media consumption analysis – run basic analyses using comScore or another planning tool to define how your audiences align to media channels.
Message testing – now that you have refined your message, test how it performs in the real world. Run a limited in-market test or test messages using a mini PPC search campaign and test multiple versions.
Attitudinal response – leverage syndicated online panels to critique your work. Validate against a control concept and verify that response is statistically different
3. Help with
in-market optimization. Once your creative is in market, the process of refinement kicks into overdrive. You need to adopt a constant process of smart ongoing creative experimentation
and refinement to stay fresh and relevant. Do the following:
Creative testing – there are three creative
elements to test: format, message and offer. Leverage experimental design to establish the appropriate test. Approaches to testing range from simple A/B to complex multivariate testing.
Audience validation – At the start of your creative ideation, you thought you knew the target audienc, but it is time to validate your assumption. Several online vendors offer audience validation tools that allow you to track how your creative properties index against your key audiences. If the audience has shifted, great! That means you have acquired net new segments (but only if you followed steps 1 & 2 in designing your creative to begin with, else it may mean you were wrong in your assumption).
By leveraging these three steps, you can create an online personalized experience supported by creative alignment to your audience needs/wants. Proper assignment of creative concepts to media channels, and constant monitoring and testing, will ensure that your creative is personalized and effective.
Michael, I am a firm believer in data-driven marketing. But I am a skeptic about data-driven creative.
I apologize in advance for the curmudgeonly tone of what I am about to say. I am *not* against data: in fact, my opinions are based on a a great deal of it.
We now have more than a decade of data that demonstrates that data-driven creative has failed to deliver the goods for brand advertisers. Yes, it has been a boon for DR marketers and e-commerce. But for brand marketers... not so much.
Click through rates remain near statistical zero. In many categories, the predicted rush of brand dollars to digital has simply not happened: if we ignore the press releases and focus on brand dollars spent, digital spending continues to languish at about the 10% level -- but this includes paid search and social media. In advertising, TV continues to command the lion's share of budgets and is perhaps the only medium left that retains any real pricing power.
I'm not saying any of this because I reject data. I'm saying this because I think the data is telling us a clear story which we are failing to heed.
I'm also not saying this because I reject digital. I founded a digital agency in the early 1990s and have been in it ever since.
I'm saying it because I worry we're continuing the same mistake David Ogilvy warned against 60 years ago: "I notice increasing reluctance on the part of marketing executives to use judgment; they are coming to rely too much on research, and they use it as a drunkard uses a lamp post -- for support, rather than for illumination."
If we really want to improve our results, we need to allow creative to *create*. We need to let it surprise us, entertain us, and move us. We also need to allow it to over-reach and fail sometimes.
There is a lot we can -- and must -- learn from data. Without it, we are blind as marketers. It is indispensable for marketing.
But we can't create magic in our spreadsheets. Wall Street has tried this and failed, as has Hollywood, as has banner advertising.
The data tells the tale: we have strangled magic with our spreadsheets. Maybe, if we relax our death grip on them, we can find a better way.
More about that here: http://www.mediabizbloggers.com/tom-cunniff/Big-Data-Big-Magic-Or-Both---Tom-Cunniff.html
Tom - I see your point and it is valid. Data driven creative is a no-brainer for DR, but it is more challenging for Brand Advertising; that is why I specifically point to social listening and search insights as starting points aligned with research.
Partnering closely with the creative team is also essential.
In the end all creative can be evaluated in-market, the result of which will lead to more of the same or a changed strategy.
Michael, thanks. I agree that in the end all creative should be evaluated in-market. A flaw in the industry's current approach to digital for brands is that we evaluate it using the wrong yardsticks: we measure "events" (CTRs, social mentions) when we should be measuring "processes" (brand preference and purchase intent over time). Also, I believe we need holistic measurement -- how does the digital effort deepen/extend/amplify the efforts in other media? We'll never have a completely satisfying answer, but -- we need to do our best.
I think this is great stuff. It's simply using technology to do things that have already been done-- but better. Top creatives have always been students of trends and pop culture, have always floated up trial-balloons...these are but a few examples of how technology allows them to do that better.
Without focus group pallor.
Michael, thanks for the article. I am working with our design team on the various qualitative and quantitative inputs, and the question they asked was when to use what input. Your framework answers that question, so I'll be leveraging it in our future discussions.
I would be happy set-up time to chat more on this topic and/or analytics in general.