sentiment analysis

Can you Listen?

by Lora Cecere on August 1, 2011 · 1 comment

This week, Altimeter hosted a pilot event at our San Mateo office.  We were sold out and enjoyed the discussion with 80 thought leaders on the evolution of social listening.  Discussion on Listening  I had the pleasure, at the event, to interviewSusan Etlinger, Analyst Altimeter (blog on her insights from the research that she completed on the topic on her soon to be released report. Here I share my insights.


Can you hear me?

Data types are proliferating. Data volumes are exploding. The certain path for Business Intelligence (BI) excellence is less certain. In the words of one of my clients last week, “We have a vision for structured data.  We buy more hardware, staff more Database Administrators and focus on higher speed processing.  We have met the challenge of 5X more transactional data in the last two years.  However, the world of unstructured data is a new frontier.  While we have scrambled to meet the structured data challenge, we are uncertain how to go about developing  vision for unstructured data. “   I think that this is our new world.

There are five elements:
Technology Worlds are Colliding.  The path to use unstructured data to listen to the customer and use it by the organization is evolving.  In he worlds of Susan Etlinger, “social analytics is aspirational.”  The interest is high, and the most frequent use cases are for brand sentiment listening, market program effectiveness evaluation and customer service.  Most of the interest is in marketing. At Altimeter, the evolution of listening platforms is one of the disruptive technologies that we are following.  The race for a solution is a collision course between two types of technologies:  enterprise expansionists and social reformers. Enterprise expansionists include companies like Clarabridge, SAS and IBM and social reformers include companies like Adobe/Ormniture, IBM/Coremetrics, Salesforce/Radian6, and Visible Technologies. We do not think that this space will be defined by what is often termed “social customer relationship management(SCRM).”  (The CRM approaches tend to be sales-driven focused on sales efficiency.  The focus is inside-out.) We feel the greatest opportunity is market-driven mining unstructured text to power outside-in processes.  The key to maximize the effectiveness is outside-in processes to listen and learn.

 The social reformers are further along on the development of data models and offer an easier to use solution.  The most common deployments are in marketing or public relations.  The enterprise expansionists solutions are deeper, more scalable and are a better fit into an enterprise BI strategy.  For the supply chain leader, these  three differences boil down to:

  1. One piece of the pie.  Market-driven processes are outside-in.  Sensing market changes to reduce the latency and improve the response.  Social is only one of many signals that offer promise in the design of market-driven processes.  Unstructured data in the supply chain is proliferating in many ways, not just in social networks.  This includes call-center data, distributor feedback, service and warranty claims, aggregated review data, B2b networks, etc.
  2. Do you know the questions to ask?  The less sophisticated social analytics platforms force the user to pre-build the questions to ask.  This assumes that the user is  knowledgable in the market to predict what is important.  If there is a market shift, the social analytic platforms are not flexible.  They cannot shift with the market.  they lack natural language processing capabilities to listen and learn and interpret market shifts.  In many ways, it is a lower level, less intelligent response. However, technologies from the social reformers are 2-5X lower cost and much easier to use.
  3. Data model.  Buy with the goal in mind.  The data models for the listening platforms are built for very different goals.  Get very clear on the intent.

Who will Win?  It is too early to tell, but we project a steep hype cycle with a long trough.  We also see convergence with large companies investing in the smaller social reformers.  The technology winner is too uncertain to predict. We feel that it will not be the traditional BI vendors or the CRM technologies.   As a result, buy the technologies with a focus on less than a two year return on investment and prepare to go through several evolutions as the market changes.  However, we feel that the real winner will be those that use the technologies to define unique differentiation.

Bottom line:  The first generation of technologies should be seen as an investment (in the words of one customer “throw away”.

Why Should Supply Chain Leaders Care?  In 90% of supply chain strategy documents that I read, companies state that they want to build a value network that extends from the customer’s customer to the supplier’s supplier.  However, the ability to listen to the customer in near real-time is new and EXCITING.  We believe that it will be a building block of outside-in processes.  It will help answer questions like:

  1. How do customers really like my new product? The technologies can reduce latency to sense true market sentiment by 80%.  With demand error of new products at 80% error, the building of outside-in processes to sense and respond could dramatically change the ability of a company to launch a new product successfully.  And, since new product obsolescence is the largest issue with product write-offs the impact could be substantial to many balance sheets.
  2. Early warning.  If there is a problem with my product in the market, how do I keep it from becoming headline news?   Today, product failures all too often become headline news. And, often history.  How many of these companies wish that they did not have this association?   Toyota and brakes.  Jeep and tire failure.  Conagra and peanut butter recalls.  Nestle and cookie dough cookie dough illness.   It is taking companies too long to determine that there is a problem and drive correction.
  3. New market opportunities. Harvesting demand insights.  The largest issue in the new product launch surveys that I have conducted for the past ten years is harnessing demand insights.  The building of intelligent listening posts can help companies gain a competitive edge.

Bottom line:  Listening posts and the building of listen and learn processes is fundamental to any supply chain focused on a growth strategy.

Who does it best?  We continue to be impressed by companies like Best Buy, Dell, Newell Rubbermaid, and Whirlpool.  Check out the pictures of the Dell listening center in Susan’s report. I love the picture!

Bottom line:  The companies that are doing it best have the strongest investments in social communities and commerce.  For leaders, it is an integral component of the strategy.
Barriers to Listening:  It is new and evolving.  Like any new technology there is a risk.  However, the largest barrier is not the risk of the technology, it is change management.  Fundamentally, organizations are incented to respond.  They are not motivated to listen.  As a result, the processes to listen, learn and drive an intelligent response are aspirational at best.

Bottom line:  How can you build demand-driven value networks without thinking about listening posts and listen & learn strategies.

Do you want to hear me?

Without investments in these technologies, if we are honest with ourselves, the answer to customer requests is “I CANNOT HEAR YOU!”  And, while historically, companies wanted to listen and to learn, today if they are not investing in the technology, it is largely because they either do not care about their customer or they are unaware.  What do you think?  Have you thought about investments to mine unstructured text to tie your supply chain to the voice of the customer?

March of the Penguins

by Lora Cecere on June 21, 2011 · 3 comments

They look alike.

Noisy and boisterous, they follow each other.  Mile after mile, through adverse conditions, they trudge.  It is a well-worn and familiar path.

When they come to the edge, they crowd together.  With extravagant gyrations they aggressively communicate, but each is afraid to take the next step.  The jump is a hard decision to make.

Last week, I felt like I was watching a re-run of a familiar movie. As I slipped into my seat at the Consumer Goods and Technology (CGT) Sales and Marketing event, I found myself in the balcony. I  looked down.  It was my seventh year of attending the CGT Sales and Marketing Conference. It was a great time to reflect back.   The event had a lot of “sameness.” It had the same themes, same people, same level of attendance, and same names of vendor  sponsors crowding the conference room foyer.  The audience looked alike –similar demographics, backgrounds and experiences–to previous years.  I value my time there and I give thanks for all the great work that CGT does for the industry, but in many ways it resembled one of the scenes in my favorite movie, “The March of the Penguins.”  Anthropomorphism in action….  <Try playing this word in scrabble.>

How so?  During the mating season, penguins gather on the ice flow and look down.  They are afraid to jump into the water due to the presence of the leopard seal. To protect themselves, they huddle together.  It is a dilemma; for, they must jump to feed and survive.  It is dangerous.  As a result, they wait for the first Penguin to jump.  They watch to see if the first into the water survives and then they all jump in mass.  Sometimes, when I am at conferences like CGT, I feel like I am watching a re-run of this movie.  These professionals know that they need to jump.  Traditional software approaches have not served them well; yet, they are afraid.  Their jobs are on the line.  They huddle to look to see who jumps.  They wait to see if the new approach works, and then they jump in mass.

Anthropomorphism in Action

Changing a ritual takes time.  The Consumer Packaged Goods (CPG) industry moves slowly.  The millions of dollars that companies have spent on multi-year projects for Enterprise Resource Planning (ERP) system Customer Relationship Management (CRM) is a painful and expensive trudge.  After studying the industry for over five years, there is no easy answer for trade promotion management, sales accounting and demand insights from these packaged solutions.  Bottomline, companies cannot build market-sensing approaches from these traditional technologies; yet, companies will not take the leap of faith to try different approaches.

I wish that the program and discussion at CGT had not been one of sameness.  The pace of change in the industry is SLOW…. Why? Product margins are high, there has been no compelling event to change, and with sales and marketing job security high, why should they take the plunge?  Why should they put their job on the line and try a new approach?  The answer is simple.  The traditional approach does not work.

Technology applications for sales and marketing are fraught with issues.  They are expensive.  As Oracle and SAP engage in hand-hand combat for trade promotion deals, the 20-40 million dollar price tag for license application deployments leaves many teams with sticker shock. The projects have a high failure rate. Based on over five years of research for CGT, three out of five TPM projects fail to meet expectations. The road to success is paved with many speed bumps that can derail even the well-intended project.

The needs of sales and marketing are also more complex.  The evolution of sales account teams and sales purchase of retail-specific applications has led to disparate applications. In interviews, even the smallest companies, have at least forty different systems distributed across the sales teams requiring maintenance, evolution and integration. The traditional Customer Relationship Management(CRM) data models are not a good fit for CPG.  As a result, Siebel (now Oracle) is a force fit and has never been a player.   SAP, on its third generation, of CRM for Trade Promotion Management(TPM) is struggling to deliver a successful project.  Accrual accounting continues to be the Achilles heel.  As a result, many grassroots efforts have spawned solutions for downstream data, deduction management, retail execution, and syndicated data.  The names of the software companies are many, the companies are small, and all are competing for attention in the  hallway outside of the CGT conference.

What should we Do?

The times they are a changing….  The manufacturer’s product margins are smaller now. Commodity’s are scarce.  Retailers have stronger brands.  They are better at analytics.  Power is shifting to the shopper.   The IT’s organizational stronghold on the organization to buy only from a standard vendor has lessened.  Software as a Service (SaaS) is a more viable alternative.  Managed services are emerging.  New approaches through social technologies now allow companies to be more customer-centric.  Yet, the scene at the CGT conference has not changed much.  The topics are the same, the people are the same, the approaches are the same, and the It is the SAMENESS that harkens the visions of the penguins standing on the iceberg flapping their wings.  I want to SAY, “JUMP Damn it!”  Spread your wings, consider new approaches.  Let’s move this STALE agenda forward.  The threat of the leopard seal in the water is far less than the market risks that are gathering on the horizon. I think that these are the new paths that we should be trudging:

How do we Sense?  Test and Learn?  Build processes from the outside-in? Last week, I had the chance to catch-up with Jim Manzi, Applied Predictive Technologies to discuss how the building of test and learn scenarios. <He is such a smart guy!> APT applies deep statistics to help companies know the true difference between correlation and causality. I feel that we would be well served to view go-to-market approaches as experiments to be tested, with rapid test and learn approaches. Recognized leaders include Family Dollar, Meijer, and WaWa.   There are also many retailers that will remain nameless–mum is the word– especially in the hills of Arkansas.  This is a major shift.  How do we build value networks to test and learn?  Today, we just respond.  Companies have fixed plans.  As markets change, they do not.  How do we use social networks like Twitter to listen better to the customer in many to many customer service networks.  I feel that this is a new path to trudge with great promise.

How do we become customer-centric?  To listen?  To Learn? To Engage in a Meaningful Dialogue, to drive Continuous Improvement? Now is the time for sentiment analysis and the use of social networks for direct dialogue with the shopper through social networks.  These approaches allow us to reduce latency on decisions, to better sense true customer sentiment and make rapid market changes.  I also think that it is time for us to directly couple downstream data with demand orchestration processes to build a horizontal platform that connects buy-side decisions (which commodities to buy when) with sell-side decisions (what to promote and how to price when) through a combination of applications like Sentiment Analysis + Price Optimization+ Downstream Data + Pattern Recognition+Risk Management to orchestrate demand.  I was excited to have my beliefs confirmed this week in the discussion with IBM Consumer Products team.  They are currently working with Relational Solutions and Signal Demand on a Software as a Service (SaaS) solution.  There will be more strategic vendors added to this road map in the future.  I believe that these are winning combinations to leapfrog the current dilemma. I also believe that there will be more SaaS combinations and managed services to emerge that combine vendor solutions that are built to help scale the current problem.d

How do we best manage global markets?  Where is the right balance of power between central planning and regional decision making? CPG companies are becoming more global.  Retailers will remain regional  Consumer products companies can now use Social Commerce platforms to disintermediate the channel.  How do companies trudge this new path and build new processes? I feel that this would have been a great discussion. For example,  Infosys has a new solution to help companies build effective demand networks with distributors in emerging economies. In my opinion, this would have been a great audience to share techniques on sharing data with distributors in emerging economies.

New ways to reach the consumer.  Tags.  Mobile applications? Exploring new channels. New technologies for retail execution.  What are the strategies for big-data value networks? Direct communication in the buying moment directly with the consumer is the new reality.  What are the strategies and how do we design these big data value networks? How do we unleash the power of mobile applications, tags, social couponing, and visualization to change the shopping experience in the store?   These are all new concepts, but do not have pat answers in the form of standard license software. Instead it requires leadership to build the path for the future.  How are others doing it?  What is the path of the future?  How are companies funding these new approaches to know the shopper?  What do the cross-functional teams look like?

Power of Reviews. Technologies that aggregate and federate review information–Bazzarvoice and PowerReviews are examples–deserve mention.  How about video reviews like Expo TV? Too few CPG companies are discussing the power of customer reviews and the strategies to federate review information with retailers to improve purchasing decisions.  Increasingly, shoppers want to hear from voices like theirs. They are tired of brands yelling the same messages.  How do we best use these new technologies to reach the new consumer.

OK.  I know it.  I have been in the industry a long time.  I am tired of standing on the ice flows talking about yesterday’s solutions.  Please, can we jump?  Can we talk about new approaches?

What do you think? Are there new approaches to solving sales and marketing problems that you think are worth mentioning?  Are you ready to jump? Please share.  Until then, I will continue to trudge–begrudgingly–with the penguins.