An Interview with Ana Athayde
Several years ago at the London International Online Show, I was wandering around looking at mostly me-too products. The content management vendors were pitching XML, connectors, and easy search functions. There were some old-line database vendors, but the offerings were mostly a rehash of research services I had used when but a wee lad. I decided on my second visit to the exhibition area that I would talk only to the vendors who offered products that were not pitching content management and for-fee databases. With a game plan, I poked around and received demonstrations from three companies. One of these struck me as having [a] a scrappy yet highly professional attitude when I asked my tough questions and [b] offered a service that moved a step beyond the "me too" pitches much loved by. CMS and other content processing vendors. Spotter pushes into new territory.
After a demonstration and conversations with two different professionals, I met the president of Spotter, a company with offices in Brussels, Dubai, Coimbra (Portugal), London, Montpellier, New York and Paris. As a dweller in rural America myself, I asked Ana Athayde, “What's with Portugal?” She smiled graciously and replied in Portuguese, “I am from Portugal.” What caught my attention was a second fact: A female starting a high-technology company in a fiercely competitive field and perfectly happy with responding to customers who have heard about her firm's content processing capabilities.
I contacted Ms. Athayde at the end of August. She agreed to an interview for the Search Wizards Speak series. The full text of our interview, conducted on August 31, 2011 appears below.
What is the core business of Spotter?
At Spotter we characterize our company as a high standard provider of Decisional Analysis. This year, 2011, is for us the year in which we will enlarge our vertical applications and keep meeting our customers’ evolving analytical needs. It is also the launch of the third version of our monitoring and analysis platform.
Are you a public company or privately owned?
We are privately owned at this time.
What is the problem that Spotter solves for its customers?
The problem that Spotter solves for its clients is to build Knowledge gleaned from information to fuel decision making. The problem solved is not only ‘Search’ or ‘Monitoring’ information, as Google can do. This is only the very first step in Analytic process: we provide Decisional Analytics.
Our clients ask for strategic input on a brand or market; not only daily alerts or monitoring features. The clients want to know more about their customers and what motivates them, about their company’s reputation, about risks in their environment; not only a search results. Our clients need deep dive analysis for decision making; not just a dashboard tool and quantitative indicators. They want to be able to interpret what it all means; not a data-dump.
Spotter offers leading edge solutions for high standard content analysis which meet customers’ analytical needs such as mapping and analyzing information about the business environment to gain strategic business advantage and making discoveries, solving complex problems and deploying results throughout the enterprise in a form that makes the information easy to use.
What pulled you to the use of technology to solve information problems?
That’s a good question. My background is business and marketing management in the sports field. In my first professional experience, I had to face major challenges in communication and marketing working for the International Olympic Committee. The amount of information published on those subjects was so huge that the first challenge was to solve the infoglut: not only to search for relevant information and build a list, but to understand opinions and assess reputation at an international level.
Other challenges were to put in place early warning systems for crisis detection and management, and in particular to analyze media perception and information propagation.
I then understood that I had to find new tools to handle the amount and complexity of information on so large a range of sources such as newswires, newspapers, television, radio, and the Web. In the past, for certain technical fields like aeronautics, nuclear energy, and finance, the management problem of large data flow was partially solved by data analytics.
I decided to fund a company to deliver a solution that could make use of information in textual form, what most people call unstructured data. But I knew that the information had to be presented in a way that a decision maker could actually use. Data dumps and row after row of numbers usually mean no one can tell what's important without spending minutes, maybe hours deciphering the outputs. So Spotter uses the knowledge gleaned from text mining to fuel decision making for a manager who is pressed for time and faces great pressure to be fully informed.
Spotter has invested 20 person years in research and development. We have built linguistic, and semantic core-technology plus the analytic solutions I mentioned. We have hired the former Datops SA (Lexis Nexis) chief technology officer Oliver Massiot. He created the LexisNexis text mining and analytics technology.
Would you describe a typical use case for your product?
Spotter offers its clients decisional analytics in several business application areas. We provide what is called “Market Sentiment” (that is, market Information, customer feedback, voice of the market, and competitive intelligence). We also provide solutions for PR and Communication Directors. This includes a more modern approach to what is called a “clipping service”.
We do public relations performance measurement so a client can get hard data about the reach of a PR campaign, and what we call Reputation Management. It makes it easy to know if a product is getting bad press or if a person is the subject of negative media coverage.
We also offer risk management. For this business we emphasize controlling risk related to certain types of public messages about a company, product, or person but also valuable legal and market information. We like to say that we can assist a customer in protecting intangible assets like the reputation of a company or being alerted about new legislation, movement of the competitors, clients and suppliers as well as sensitive issues in their market sector.
And we provide strategic information services. Our Spotter technology makes it easy to know the market environment, understand shifts in the world economic picture, and adjust particular sales and marketing tactics in accordance with the information Spotter delivers.
Can you give me some concrete examples?p>Yes, of course. A typical use case of our market sentiment solution is for a business in the fast-food industry field, for example. Our client can monitor online consumer buzz in social media. The customer can also evaluate communication opportunities. Our service makes it easy to investigate and understand determinants of consumer satisfaction and dissatisfaction with key attributes of products and brands such as usage, prices, and image. We provide then to this client a full solution from information monitoring to decision.
What does the customer get from your firm?
A typical client engagement delivers several components. First, there is tracking and monitoring media and social media. Our technology indexes a wide range of social media content and can integrate data from various providers. We apply what we call specific structure surveillance; as part of content processing. These algorithms have been adapted to each type of media content such as articles, postings on a blog, comments, discussion threads, tweets and so on.
Second, our linguistic and semantic engine applies text mining and sentiment analysis techniques. For us, this means entity detection, speech detection, and sentiment analysis. When we perform consumer conversation analysis, our text mining and indexing systems will extract the “who says what and how on which product or brand’. In this process, we deliver to the users a &ldquot;monitoring dashboard”. The customer can see topics, sources, products and brands sentiment scores. One useful feature, according to our customers is one-click access to the most relevant conversation, the most negative versus positive. I think you call this “search without search”. Our clients don't want to search. The clients want information to know a situation quickly. Let me show you.
Third, we provide what we call “business oriented models.”; Our proprietary analytics engine will then apply business rules and metrics for a better understanding and assessment of influences. These models deliver actionable information on the potential impact on global consumer opinion or a forecast of certain trends before these appear in the New York Times or a competitor's blog. Our graphical dashboards and reports are customized for each group of users in a client organization. These “dashboards” provide overviews or deep dive analysis adapted to the organization decision-making processes.
And, finally, we provide what we call a “decision management system”. This component helps one of our licensees to create strategies. The decision management system can also provide a foundation for the measurement of the impact of online and offline communication. The system makes it easy to fine tune a mix or sales and marketing methods. The idea is to prevent loss of online reputation and opportunities to manage social media messaging.
What differentiates your service from other commercial services?
Yes, that’s an important question. There are main differences between Spotter's analytic solutions and most of the services available today. We talk about four key points.
First, relevancy: a simple query gives some fuzzy results or none business focused information. On an example such as a query on Fortis Bank NV, most of the stories on a competitive service are documents which don’t contain any information on Fortis, but are papers where analysts from Fortis offers opinions. Spotter’s dashboards rely on an expertise on business activity, challenges and issues, on advanced text mining technologies and analytic metrics and models on which confidence is guaranteed.
Second, professional versus non-professional use. Some of the content processing systems are aimed at the casual user. We focus on business professionals who know quite a bit about a particular field or business sector. We tell our potential customers that Spotter's solutions are dedicated to the professional users. For example, the content of a dashboard’s sources and topics among other facets is discussed and customized for each customer. The filters, the metrics and the results are tested and validated.
Third, analytics: Most of the business intelligence systems provide basic quantitative graphic outputs like a bar chart. Spotter solutions rely on semantic analytics such as sentiment analysis which are the basis of consumer satisfaction assessment and risk rating and we also employ a business rules engine which enables real qualitative deep analysis of the processed content.
Lastly, we have broad language support. Most services work on Spanish, German and French content, but the bias is toward English and English language sources except in specific situations. More than 95 percent of Spotter’s clients need international monitoring and surveys. Spotter provides solutions which handle any languages in the whole monitoring, indexing and analytic processes we provide.
Are you running your own crawlers, or are you a metasearch service?
Spotter’s tracking and monitoring system can combine different crawling strategies. For example, many of our clients want us to syndicate results from information aggregators and databases such as Factiva, LexisNexis, professional specialized databases, internal databases. We also accommodate content the client may license directly as well as particular sources that the client wants to process such as a list of Web sites and blogs.
What's the architecture of your system?
The architecture of our own crawling system is based on proprietary methods to define and tune search scenarios. The “plumbing” is a fully scalable architecture which distributes tasks to schedulers. The content is processed, and we syndicate results.
We use what we call “a source monitoring approach” which makes use of standard Web scraping methods. However, we have developed our own methods to adjust the scraping technology to each source in order to search all available documents. We extract metadata and relevant content from each page or content object.
Only documents which have been assessed as fresh are processed and provided to users. This assessment is done by a proprietary algorithm based on rules involving such factors as [a] the publication date. This means that each document collected by Spotter’s tracking and monitoring system is stamped with a publication date. This date is extracted by the Web scraping technology, from the document content. [b] The type of behavior of the source; that is, the source has a known update cycle. [c] We analyze the text content of the document. And [d] we use the date and time stamp on the document itself.
I understand that you offer a full service package. Aren’t you targeting a very crowded market?
I think the key point we try to make clear is our “bundle”; that is, we deliver a solution, not a collection of puzzle pieces. Our approach is truly international. Our ability to capture, monitor and analyze decisions and their impact requires rich, higher order meta data constructs. Many companies such as Autonomy, Microsoft, and Oracle also promise similar services. But once this has been done, the process of information toward decision is not complete. The main competitive advantage of Spotter is to be able to provide to its clients a full decision-making solution which includes, as I mentioned, analytics and our decision management system.
Can you give me an example?
Sure, let me give you a snapshot about the depth of functions we provide to a customer. A marketing person can get overviews or do a “deep dive” analysis. We offer specific analytics packages for risk management. The manager can access customized models for risk assessment, perform risk scoring, and generate reports without coding or digging through FAQ files. Add to this our complete cycle approach from information monitoring to the making of a decision. Our solution is engineered to link efficiency and quality control throughout the content processing “chain.”
When you look forward to 2012, what are some of the new features you and your team hope to introduce?
We are thinking about two enhancements to our system. First we will be extending the analytic tool box with advanced visualization capabilities. Some questions can be answered with visual outputs and for predictive work, a visualization is one way to spot a trend.
And, second, we are now packaging additional sets of business rules and metrics for specific business functions. For example, we want to offer a marketing customer a package that is "tuned" to the needs of that specific business function. We are almost completed with a packaged solution for the risk management industry as well.
Our focus is to give the client flexibility and at the same time deliver a solution that can be deployed and used very quickly, maybe in one day or less.
What systems do you support?
Our system can be used with almost any organization's existing infrastructure and systems. We provide a Web-based interface and an application programming interface for gateway development for proprietary enterprise systems.
So you support SharePoint?
Yes. We already compatible with most enterprise software and systems. If we find something for which we don't have support, we add it to our system.
When you look forward to 2012, what are the main trends you see taking shape in search and content processing?
Content processing is, of course, a fast-moving business area. We have one advantage, of course. We use Spotter to help us determine where the market is heading. On our list of trends is the entire social media sector. This is of interest to many of our customers at this time. Some products can be helped or hurt by information in the social network sector.
I also think that there is more interest in a combination of text and numbers when researching a topic. I think of this as a "text numeric convergence." I know we are working now to make text mining functions provide the same level of meta-information that data-mining already uses in traditional business intelligence systems. In fact, text mining methods are being extended in some interesting ways; for example, sophisticated extraction technologies and vertical market data mining applications.
We are cognizant of the growing interest in sentiment analysis within comments, blog posts, tweets, opinions, and latent consumer attitudes.
One final question, would you encourage other women to jump into the search and content processing business, or is the "glass ceiling" too tough to penetrate?
There is no real specific barrier for a woman to jump in this business even if challenges are high to handle in term of complexity and diversity of the assets you need to build. In my opinion, there aren't today many women in this business, and we would welcome others; so of course we encourage women in this area of expertise. We value diversity, knowledge, competence, equal rights between men and women. Our team includes many women and their skills go from mathematics, multi languages to multi cultural environment.
Where can a reader get more information about Spotter?
I suggest a visit to our Web site at www.spotter.com.
Spotter has generated considerable interest with its "ready to roll" approach to content monitoring and business intelligence. Unlike some of the companies in the content processing sector, Spotter has a full line up of components that deliver a range of business intelligence and decision support services. At this time, most of the firm's business is in Europe, but there are signals that Spotter will be expanding its presence in the North American market in 2012. For more information about the company, navigate to www.spotter.com.
Stephen E. Arnold, September 6, 2011