During the past month, I was able to attend two premier conferences dedicated to predictive analytics; these were the The Predictive Analytics Summit in San Diego, and Predictive Analytics World in San Francisco. In categories like this, it’s next to impossible to follow the notion of a mutually exclusive and exhaustive set. In spite of this classification issue, I will do my best to provide a bird’s eye view of the sector. I’ve broken it down into three types of predictive analytics and the three types of companies pursuing the predictive analytics market.
Latest Entries »
We often hear how product failed in the market, particularly when company launched on ‘gut’ feelings. But there are several cases where market research was integrated and utilized as strategic partner which eventually resulted into great brand success. One such case is Hyundai – below is the case study on how Hyundai used market research to identify unmeet needs, to understand latent behaviours, and drivers. An excellent case of using quasi-quant approach. This success was also propelled by senior management’s trust on market research.
Issue: Americans were laughing at Hyundai’s cars when John Krafcik joined the company eight years ago. The cars were ugly and often broke down. The only reason to buy one was because it was cheap. Comedian Jay Leno once joked that you could double a Hyundai’s value by filling it up with gas.
Senior management believed that first they had to build that foundation of quality and consumer trust. By the time Hyundai to the late ’90s, they already knew that their quality was good. Hyundai knew their reputation was horrible. And the America’s Best Warranty, that 10-year, 100,000-mile powertrain warranty, ended up being huge. Hyundai were able to take themselves from about 90,000 units (cars and trucks) per year at the absolute pits of our sales, this was in the late 90s, to 300,000 and 400,000 within four or five years.
Approach: Based on market research, Hyundai realized that midsize cars should be styled conservatively. Products starting with the `05 Tucson and the `06 Sonata were designed in a safe and conservative manner. It didn’t move the needle in terms of sales BUT Hyundai finally saw the pattern. So now when Hyundai does research, they categorize people before they come (in) as either conservative or progressive in their design thought. And Hyundai can then weight the results. It gives more weight to what Hyundai call design progressives.
Additionally, Hyundai extensively used observation research. Hyundai sends a couple of product planners and market researchers out with consumers and just literally spend a day or two with them, looking at their cars. Seeing how they live their lives around their cars. When they go to Costco, where do they put those big, huge things of toilet paper? It gave tremendous insights especially surfacing unmet needs.
Result: The Korean automaker’s quality has improved, and it’s among the leaders in fuel efficiency and styling. Sales are up more than 60 percent since 2008, the year Krafcik (pronounced KRAF-chick) became CEO of American operations. Hyundai’s Elantra compacts and Sonata midsize sedans are in such demand that few discounts are offered. And although the company’s U.S. sales are just a fraction of General Motors’ or Ford’s, they’re growing so quickly that Hyundai is feared by every other carmaker.
Hyundai had already started to change before Krafcik arrived, offering a 10-year, 100,000-mile (160,000-kilometer) warranty with its cars. But the transformation from joke to juggernaut accelerated under his watch.
Source: Economic Times
Comparing Google Consumer Surveys to Existing Probability and Non-Probability Based Internet Surveys
Google enters into market research claiming superior sampling. This new tool not only offers data collection but automated top-level analysis including statistical significance. I highly recommend fellow researchers to read google’s white paper on sample representation. Below is the abstract:
This study compares the responses of a probability based Internet panel, a non-probability based Internet panel and Google Consumer Surveys against several media consumption and health benchmarks. The Consumer Surveys results were found to be more accurate than both the probability and non-probability based Internet panels in three separate measures: average absolute error (distance from the benchmark), largest absolute error, and percent of responses within 3.5 percentage points of the benchmark. These results suggest that despite differences in survey methodology, Consumer Surveys can be used in place of more traditional Internet based panels without sacrificing accuracy.
White paper: http://www.google.com/insights/consumersurveys/static/357957729093014734/consumer_surveys_whitepaper.pdf
So finally, Google enters into market research as data provider. What impact you think this would have in MR industry?
From sawtooth:
One of the more common optimization approaches in market research is TURF analysis. TURF stands for Total Unduplicated Reach and Frequency. It is a technique for finding the optimal bundle of items (messages, flavors, brands, magazines, etc.) that “reach” respondents. TURF has been available for MaxDiff problems using our Online MaxDiff Analyzer tool. But, we’ve now made that same tool handle general data other than MaxDiff including Likert scale, ratings, magazine readership (yes/no), brand within consideration set (yes/no), etc.
A classic TURF example is the problem of choosing which flavors of ice cream to stock in the freezer at a grocery store. The grocer may decide that he/she has limited space and can only include up to 8 flavors of ice cream (out of 30 possible flavors). The grocer wants to maximize the chance that shoppers will find a flavor that they like well enough to buy in the freezer (where “like well enough to buy” means that the preference score exceeds some threshold, such as a 3 on a 5-point scale). If the respondent encounters a flavor he/she likes in the freezer (having a score of 4 or 5), the respondent is counted as “reached.” The problem isn’t as simple as including the eight most preferred flavors on average across the sample. Niche flavors that appeal to segments of the population (and that can increase total reach) would be overlooked.
For the ice cream example outlined above, the TURF procedure examines all possible subsets of 8 flavors of ice cream (out of 30 total flavors), and for each set counts how many respondents are “reached.” The top sets of 8 flavors that maximize “reach” are listed in the output with the percent of respondents reached shown next to each.
The MaxDiff Analyzer software is a SaaS (Software as a Service) solution, meaning you don’t have to install anything. You log on to the software using a browser. The cost is just $250 per year, for access to up to 10 projects concurrently. Read more about MaxDiff Analyzer by visiting: http://www.sawtoothsoftware.com/products/maxdiff/maxdiff_analyzer
Much that researchers and clients desire to utilize social media, there still remains few skepticism: representativeness, people aren’t talking about my area of interest; metrics aren’t accurate or suggests what I already know etc. However, there has been significant progress. Here are my top picks of articles I found insightful. These articles may help you to better understand use, implication, and ROI of social media. Selection is slightly biased to pharmaceutical industry because of my background 
. Do share other tips or articles that you find interesting.
- Using social media research to complement traditional methods – Quirks
- What happens when you mix panel respondents and social network respondents? – Quirks
- Market Research: Tapping into New Market Research Opportunities – Pharma Voice
- Customer Centric Segmentation – Pharma Voice
- KOL Mapping: The GPS of Thought Leader Identification – Pharma Voice
- Multichannel Measurement: It’s Not All About ROI – Pharma Voice
- Hitting the Social Comfort Zone – Pharma Voice
- Healthcare And Social Media: A Winning Formula- Mckinsey and Company
- Three ‘Game-Changing’ Ways of Using Social Media in Healthcare Market Research – PMRG
- What Qualitative Researchers Can Learn from Facebook. – American Marketing Association
In the next newsletter, I will be summarizing each of the above article including surfacing key learnings. Should you wish to receive this, please subscribe.
It is well known fact that marketers’ half of the investment yield unfavorable return but the challenge is to identify which half is not working. Marketers turn to market research hoping to find information that might help to fill the black box. We researchers do reasonably good job in generating usable information investigating unobservable constructs. After all, isn’t the whole purpose of research to unearth unobservable facts? Yes, true, but the purpose was to fill the black box and tie it back to observable facts. In order to generate an actionable ‘solution’ and ‘direction,’ market research should produce evidence based summary that explains not only the attitudinal implications but also financial implications. Research should be able to address the financial metrics the ones on which marketers are measured. As such, research needs to speak financial language to help marketers make calculated and measurable risks and decisions.
With this objective I will be sharing few insights, know-how, and tool and techniques in three part series. The series will focus on how to integrate financial data into research and how to tie research with financial marketing metrics. This will also include on how to use financial data to strengthen research recommendations based on financial rationales. The information will be based on my experience and interpretation of several related articles, journals, and books by thought leaders.
Stay tuned…..
To get an auto-alert on this topic, click on subscribe button
Rather than grow larger and larger panels with dozens of profiling questions repeated over and over, perhaps we should find new ways of identifying qualified participants. With some very basic filtering and analysis of my customer database, for example, I can reach response rates in excess of 30%. By targeting followers of relevant brands on Twitter, I can do the same in those cases where I need to reach outside my CRM system.
via Step 2: Surveys Arent Your Only Quantitative Data Source.
I am excited to summarize a presentation on a topic which have been discussed for ages but still has same vanilla flavor with rarely new toppings: ATU brand tracker. I am summarizing a research conducted by Kantar health on agency-client perception and tug-of-war on ATU studies.
If you are healthcare researcher, you already know that ATU projects are working capital generators – high in volume but modest in revenue and profitability. However, I find ATU tracker very challenging because clients tend to expect “wow” factor/insights wave-over-wave. Ring a bell of having similar situation? You are not alone; let’s see the results:
Let me begin by presenting and opportunistic and depressing fact: 50% of clients are unsatisfied with current reporting. In client’s words “I am satisfied. But I don’t expect them to add value…”
Client-agency engagement gap:
The survey evaluated feedback on several initiatives including:
Basic charting of all data + tabulations • Key metrics available via online dashboard • Specific client-requested analytics via hourly consulting model • 100 Hours of consulting time included • Additional hours available • Cost savings of approx. 15% v. full-service model
However, the study concluded that transformation innovation is “nice-to-have” but might not significantly drive client satisfaction. Clients are still looking for insights – a real time insight wave-over-wave but with same or reduced cost.
MY THOUGHTS: The results only attest what we already know. Considering drying pipelines and multiple products on their last patent journey, budgets are and will continue to shrink. Needless to say, market research continuous to be under axe. As such, researchers needs to consistently work towards client satisfaction while managing bottom-line. For an agency, ATU studies are great source of committed yearly revenue. However, lower margins, increased efforts, particularly fulfilling ad-hoc and follow-up request make it financially less favorable. As such, efforts should be strategically invested. Considering the frequency of deliverables and pressing timelines, agency might not need to invest in innovate methodologies but on how to use existing methodologies in more refined and targeted manner. Agency don’t need to do different things but do same thing differently. Below are few pointers that might help to inch up client satisfaction on ATU deliverables:
- Integrate brand plan imperatives in the report. For e.g., insert brand awareness, usage targets and directions (derived from brand plan) into ATU slides.
- Integrate prescription data to identify impact of shifting awareness and attitudes on Nrx or Trx
- Provide awareness and attitude simulator (let me know if you need to know more about this)
- Present the findings in visually appealing format – I am a big fan of key driver analysis and perceptual maps
- Write straightforward, short, direct, and targeted top boxes and executive summary
- Don’t provide 50+ slider deck – focus on quality than quantity. Data heavy reports are sure-fire way to have your report dumped in shelves. If slide do not show a trend break or significant movement, push it to appendix.
Annie – Social Media Research and Analytics
Green Book Market Research Blog
Next Gen Market Research – Tom Anderson
Quirks Market Research Blog
Voice Power – Lynn Woodman