Big Data – Framework to Be Smarter with Analysis

Marketers use data to assess campaign success, allocate budget, delineate customer analytics, design new products, optimize acquisition and retention, etc. However, with multiple data sources and owners, organizations often struggle with consolidating the right data and reporting to fully answer the questions asked. Along comes big data!  

Big data is exciting and challenging at the same time, since it affords a holistic look of customer analytics, yet exacerbates the complexity due to legacy data-silos and stakeholder needs.

This is a call for "smarter” approach to defining and using big data for tactical and strategic marketing objectives.

For marketers, a mapping exercise, as below, may help create a framework of available and needed data and systems. The layout identifies the evolution of data capture and applications, as organizations mature, and highlights the value being injected back into the business. It is telling in different ways – gaps in data capture, systems limitations, analytical resource allocation, etc. But most importantly, it highlights the need to innovate and upgrade legacy systems such that the analytics efforts can be targeted for the greatest impact and ROI.

It must be noted that the data silos are identified as they relate to customer analytics. Note that, as a marketer, I am looking into our current systems, so we can model returns from existing capabilities and tie the improvements to enhanced future investments. In other words, try and look at the parts to solve for the whole - get some short-term wins to articulate investments and the need to be on an "analytics fast lane.”

The question to ask of the management is – here’s where we are and here’s where we can be. Are we ready to invest?

The framework may help articulate big data scope and ROI, and also facilitate:

• Breaking down projects into chunks of easily manageable proposals
• Identifying the next big investment 
• Developing internal capabilities (organizational and personnel) 
• Providing a launching pad for execs to commit to larger projects

With investment along X-axis, we can improve our capabilities along Y-axis, which in turn, contributes to the growth trajectory graph. 

Understanding a Survey Design – Lessons from the Presidential Polls!

In the politically charged environment, I noticed that surveys are often used as a tool (arguably) to form public opinion. The research and analysis is inspiring and made me wonder if the insights could be directly used towards marketing effectiveness?

A recent HBR article stated that polls dramatically over-estimate support for third-party candidates, when very little actually exists. The premise being “naming any third-party candidate” will garner more votes from responders who are on the fence about Clinton or Trump. However, it won’t translate into real ballot due to the lack of familiarity with these candidates. The revealing conclusion was that the independent candidates cannot be overly optimistic about the survey results, since being named in the survey will alone get you some votes, but the actual support may be much less than that.

Now let’s think marketing surveys – most tend to be “leading” the respondents, so we as marketers can get specific data that we “like.” Not a good sign, if you are banking on survey results to build the business case for a new product/service and have been leading the consumers into picking an answer on the survey. The responses will point to what’s listed on the survey, instead of providing an insight into how much do the consumers know about and/or need the new offering.

Another related and interesting reference in the above article was to a book on “Answering questions versus revealing preferences,” by Zallar and Feldman. The research argued that responders are not necessarily revealing “considered opinions, instead are just answering questions quickly.” The responses do shift based on the answers choices given!  

If I were to summarize in marketing context;

“Educate and NOT lead the consumer to get a more considered opinion.”

This is an important consideration, when designing the survey and considering the responses. Consumers are not sitting around thinking about our new product or service, so a “considered opinion” is the last thing we should expect. However, we can still strive to help them provide a considered opinion through a better survey design, in terms of educating the consumer about the purpose, product/service, benefits and goals. How we do that will depend upon the specified goals, but the results will be closer to a considered opinion, rather than a multiple choice selector.

The article brought to fore some pertinent nuances, even while drawing attention to more common approaches employed, such as, be specific and short, or squeeze in a few more questions to get more responses. But, it is up to us to decide if we want “more” data or “quality” data. The choice is ours!

Are You Ready For the Anlaytics Fast-Lane?

Data and analytics are everywhere, with numerous examples of how the right analysis yields significant lifts in marketing or operational efficiencies. The returns are, without doubt, measurable and worth the investment. However, a word of caution for the leaders – don’t be swayed by case studies and peer recommendations, or assume that employing an analytics firm will yield similar results for your business. Being ready to utilize the feedback in a timely manner, with a clear implementation plan, is equally critical to realizing the ROI on your analytics investment. Here are a few questions to ask yourself, before embarking on an analytics project.

1. What is the problem statement? Be clear with what question you need answered. It could be; defining the target audience for your product or/service, understanding customer engagement with your product, improving acquisition effectiveness with better offers, identifying churn propensity by customer cohorts, etc. The key is to explicitly state the goal, stay focused and avoid the “noise.” Defining a clear problem statement upfront is crucial to staying focused and not be distracted by a lot "interesting" findings that are bound to pop-up.
2. Do you have the right data set? Work with your analytics and IT experts to identify the right metrics needed to answer the above questions. A third party perspective is always recommended in such instances, since it helps question the status quo and is not bound by what is familiar. Start as comprehensive as possible, since analytical modeling often throws new data dependencies that may not have been obvious. A holistic view of data points available from the data warehouse go a long way in defining the problem statement.
3. Is the organization ready to ingest the analysis? The best time to use the analysis is “now.” I have often contended that analysis based on historical data is like playing catch-up. But, with predictive modeling we can project certain behaviors with a fair degree of certainty. The imperative hence is, that an organization has the operational capability to act quickly on the recommendations (marketing & sales changes, product updates, online experience edits, etc.). Invest in the back-end systems that can adapt and learn from the new programs, or else, run the risk of being obsolete.
4. Do you have dedicated personnel to guide the process? This is the most important determinant of success, and perhaps the most overlooked as well. A well defined problem statement, predictive analytical models and process efficiency cannot be achieved unless we have the right analytical minds leading and nurturing the program. As an organization, we need to recognize the need for analytics leader who has the resources and can rally the operational teams to achieve the desired outcomes. The ROI of analytical projects depends on this critical investment, just as it does on problem statement and analytical modeling.

Analytics and data modeling empower the businesses, and to stay competitive, businesses need to equally weigh continuous innovation and implementation. Rapid deployment is as critical for success, as is harnessing and modeling business metrics!

Marketers - Re-assess Your Marketing Plan Now!

With a quarter gone, marketers ought to be digging into business performance and results. The lessons learnt and course-correction (if needed) should be clearly outlined for the remainder of the year. While the lessons are one for the books, the insights about course-correction dictate how we end the year. And, irrespective of how deep we are into our marketing commitments for Q2, I have found that this is a crucial time to revisit the marketing plan. Slow down after the frenetic pace of Q1 (as it often is in many industries), and spend some time to analyze and level-set the expectations for the rest of the year. Yes, I am suggesting slow down – just think of the benefits:

1. You will most certainly have a better sense of any corporate-level tactical shifts, and their likely impact on revenue projections and upcoming marketing campaigns
2. You will likely learn something new about your product performance and customer-buying propensities
3. You will be able to devise tactics to ensure revenue commitments stay on-course, based on 1 & 2 above
4. You will end the year with a bang – guaranteed!

So, what are we to consider now? 

• Analytics – make sure Q1 results are in and being analyzed for ROI by product, customer segment and marketing channel, not only for attribution, but also for tactical moves in Q2 and Q3. What is going to be the marketing goal based on these results? Is it better engagement with customers, or drive more sales, or brand promotion, or product updates? The answers to these will greatly impact your acquisition and retention models for 2016!
• Seasonality – you probably have a good grasp of the seasonality in your industry, sales, etc. But, make sure that there are no extraneous factors that may warrant changes to your campaign strategy. For example, consumer-spending tends go into a lull in Q2, before picking up again in Back-to-School and Holiday seasons. How is it being projected to be this year for your line of business?
• Customer Targeting – with slower months, it is also imperative that, to maintain a healthy ROIs and CPAs, more valuable leads are targeted with relevant offers. Sharpen your mailing list and tailor the products to the segments that are more likely to buy now. Predictive acquisition and retention models ought to be able with better campaign design and retention efforts!
• Communication Channel by Product – In my multi-product environment, I always look at what products are more likely to appeal to a buyer, through which channel. It could be based on demographics, geography, interest, etc., but that email, or content marketing, or newsletter that is in the works, better be "more" relevant. For example, online Travel used to take precedence over online Shopping during these months. What makes sense to sell to your target customers during the lean months?
• Budgets – understand the budget spends and make adjustments based on any new corporate imperatives. Usually, any over- or under-spending in Q1 could be corrected quickly during this phase. The ROIs and attribution by channel will shed more light on how is each channel performing for the brand.

A little postmortem of marketing performance from Q1 is probably one of the more important projects that are often inadequately addressed or staffed. If done right, it instills discipline and focus, which in turn, promotes efficiency and effectiveness of marketing operations. As I said above, we want to end the year with a bang!

Making sense of Twitter data

Marketers are trying very hard to make sense of the twitter data, but the range of "emotions" conveyed by the tweets and the lack of quantitative framework, make it extremely difficult to measure and apply the insights into real-life business improvements.

But, there are cool new ways evolving and recently, Harvard Business review had an interesting blog on "four ways of looking at twitter." Here's the link to the full article (http://blogs.hbr.org/research/2010/02/visualizing-twitter.html). The author talks about a couple of new visual aids (ex., Twitter Venn diagrams) being developed that may provide some quantitative feel to the tweets!

I decided to test my theory on "lack of analytics in strategic, operational and ecommerce decision-making" by running the Twitter Venn for the terms strategy, analytics and ecommerce. Below is the Twitter Venn that comes up:



This is a bit surprising even to me - the frequency of analytics and ecommerce tweets is relatively small, but more telling is the lack of overlap in analytics/strategy/ecommerce. I have worked with businesses where double-digit gains are easily achieved by merging the above three concepts. The tweeters, with a stake in strategy, analytics and ecommerce, just need to be thinking more congruence and overlap to realize those benefits.


Incidentally, I have been helping businesses build sustainable growth models by applying analytics to ecommerce strategy and planning (http://www.wsapartners.com for more info).

A new approach in 2010

As we head into the New Year, there will be fresh focus on the business outlook than what happened in 2009. A lot of questions need to be addressed first-up, before we get buried in the daily operations - did we achieve the most out of our business analytics efforts in 2009, was the data adequately supportive of our business strategy in 2009, do we have a clear roadmap/plan for 2010, and what relevant metrics and analytics focus in 2010 will help us achieve corporate goals, etc?

Let’s take a moment to know our strategic focus, be it at the corporate level, business unit level, or simply for the team. If we are an analytics shared service team, we need to know the key metrics that will help our business achieve the desired results. If we are a functional team, we need to assess our metrics and sit down with our shared service folks to understand the metrics and outline a process to measure/report/analyze these in the coming months. And if you are a do-it-yourself organization - well, you are over-worked - yet, you need to ensure that your functional success is aligned with analytics goals.

So where should the focus be?

E-Commerce sites may need to look at conversions, CPC rates, click-through rates, etc. Content push sites may need to look at programming efficiencies, click-through rates, repeat visitation, etc. If your goals are a blend of the above, conversions and click-through rates may be more important. Do you have an online video strategy and are the metrics appropriately defined and tracked? What about your social media strategy – are the metrics defined and tracked? Oh, and don’t forget the consumer experience metrics – ease of navigation, bounce rates, engagement depth, etc. will be great success indicators, while A/B testing a great tool to continuously upgrade your site’s experience.

The key will be to understand the business strategy for the year and outline the metrics and analytics approach at the onset – meeting with your leader/executive and setting those benchmarks is very helpful in the beginning of the year and brings focus that helps react to business contingencies. As an example, in 2009, I helped an organization stay on course to beat their OIBDA targets, in spite of the downturn in revenue.

2010 is bound to present much greater opportunities!

Where does conversion suffer?

Only 40% of companies are applying some kind of testing AND targeting to grow consumer engagement and enhance conversion. The technology is often in-house and to top it all, most companies are not sure about what to test, how to test, what metrics to test, what is the timely reporting and format needed and how to apply analytics to the data.

These findings from a survey by Omniture, reveals a lot about the problems we face in the real world. To maximize the return on the online presence, we need to define the testing and targeting strategy, and apply analytics rigor to better understand our consumers’ needs and create opportunities for business growth.

Behaviorally targeted pages with dynamic content result in a 12% lift in conversions, according to the same study – that will be quite a handsome reward for our 2010 online efforts!

Challenges with using web analytics in marketing!

In a recent study by eMarketer, top 3 challenges with using any web analytics data include; inability to integrate data with other marketing solutions (46%), verifying accuracy (41%), lack of comprehensive data (32%). Not surprising at all!

A couple of theories jump out of these findings. One, there is a lot of data that can be available through simple tracking with any of the wide array of tools in the market.

Two, not enough effort and foresight goes behind defining the success metrics and the data that needs to be collected.

Here, I want to address the second theory, since we can and should control it and demand more from it. The data needs for each marketing campaign fall into a tightly linked planning, launch and post-launch stages. If we are to take a guess, this is usually not the case!

It is imperative that the campaign strategy clearly defines success metrics, data needed to verify this success, and a clear implementation plan to ensure appropriate reporting will be available to analyse the performance. It should be made a part of the project planning process, with the campaign manager directly accountable for reporting on campaign success.

Professionally, I have been witness to several projects, where measurement and analysis could not be completed, because analytics department/individuals were not engaged from the start of the campaign process. Effective ROI measurement and future marketing investments are dependent on how robust our data quality is - that should be justification enough for engaging appropriate level of analytical and data support at the concept stage of each marketing campaign.