Read Analytics and Data Analysis Blog News

An organization may be at one of five analytical competition stages according to Angoss Software. Analytically Impaired organizations must get accurate data to improve operations. A company at the Localized Analytics stage require analytics to improve one or more functional activities. Organizations with Analytical Aspirations are ready to use analytics to improve capability to act on analytical results. Companies at each of these three stages are cloud computing candidates who may be willing to outsource and begin using analytics as a competitive advantage. Analytical Companies and Analytical Competitors include those that are prepared to make significant technology and human capital investments to expand analytics to their business users and master analytics as a competitive advantage.

Social Media and Text Analytics in the Cloud discusses how Angoss Software can help an organization implement cloud computing, predictive and business intelligence analytics as a competitive advantage. Cloud Computing Analytics and Data Encryption discusses how data encryption technology allows Angoss to work with clients so that data may be analyzed but protected. Angoss recently answered questions about the five stages of analytical competition. It also addressed how Angoss may help an organization implement predictive and cloud computing solutions analytics as a competitive advantage. The impact of social media and social network analytics on the field of predictive analytics is also discussed below.

Organizational Analytics Readiness

Question 1: How does an organization know whether it is ready to invest in Angoss’ analytics products and services? How does Angoss help with this?

Angoss: Angoss works with organizations at various levels of readiness for predictive analytics, and we have an offering for each level. For organizations without an internal capability, Angoss offers KnowledgeCLOUD solutions: which combine technology, client management, modeling, and consulting as part of an ongoing subscription service. For organizations just starting out, Angoss KnowledgeSTUDIO and KnowledgeSEEKER software are easy to use, with intelligent wizards and no coding language to learn. We have a very strong decision tree and strategy tree (for deployment) offering, and these are models that are easy to grasp and lead to early successes.

Organizations come to Angoss for many reasons. Predictive analytics is seen as the next logical step after an organization has its data under control and invested heavily in data warehousing and business intelligence. Competitive and other external pressures mean that many organizations sense that predictive analytics is something they should be doing. Sometimes, there are amazing internal champions that are carrying out a vision for their organization. As part of the Angoss sales process, we spend a lot of time probing where an organization is in terms of readiness so that we can make the best analytics recommendations.

Organizational Analytics Needs

Question 2: The Angoss website mentions organizational ‘needs’ that your analytics software can meet: sales and marketing and organizational risk. Please describe these needs and maybe mention the particular Angoss software that is best for each.

Angoss: Angoss’ software offering consists of two products: KnowledgeSTUDIO and KnowledgeSEEKER. KnowledgeSTUDIO contains everything in KnowledgeSEEKER, but it includes advanced modeling algorithms. StrategyBUILDER has been folded into both of these products.

Organizational risk – Many Angoss clients use our analytics software for credit scoring (consumer credit risk). In this case, they prefer KnowledgeSTUDIO’s scorecard module.

Angoss Analytics Case Studies

Question 3: There are many excellent Angoss analytics case studies and white papers on the http://www.angoss.com/ website, but can you briefly describe one that best communicates how Angoss’ analytical software and solutions may help an organization?

Angoss: In addition to case studies and white papers, Angoss has been building a library of videos online. I’m not sure if there is one best one. I would have a look at the KnowledgeSTUDIO Quicktour for a software overview and mini demonstration. The recent white paper on market basket analysis is good, and there is an accompanying video on the videos-software page.

Social Media and Analytics

Question 4: Has the proliferation of social media (eg. Facebook, Twitter), smartphones, and web-based applications had an impact on the analytics process? Which Angoss analytical techniques (eg. decision tree, market basket analysis, logistic regression) are the most useful?

Angoss: This depends on the problem you are trying to solve. Decision trees and logistic regression are useful when you have a goal in mind (eg. ‘what is the profile of a Facebook user who shows positive sentiment towards Angoss?’). Clustering is useful if you are trying to find homogenous groups or segments. So, maybe you are trying cluster Facebook users into a manageable number of groups that you can then deploy strategies against. Market basket analysis is about associations. For example, we could do text analysis on all the tweets containing the word ‘Angoss’ and extract the entities and themes. Then we could do market basket analysis to see the relationship between those entities and themes. Our market basket functionality includes an ‘item attraction map’ to visualize those relationships. So, for tweets mentioning ‘big data’ are they also likely to mention ‘Teradata’, or are they more likely to mention other database platforms?

Analytics Success Measures

Question 5: Has the proliferation of social media (eg. Facebook, Twitter), smartphones, and web-based applications had an impact on the analytics process? For example, the success metrics that must be used for evaluating a model once its results have been acted upon?

Angoss: The proliferation of social media and other data sources has not fundamentally changed the success criteria of a model deployment. If a model has been deployed, then ideally you have set up a randomized experimental design so that you can compare decisions based on the model with decisions made without the model. So, a lot of the evaluation is doing that comparison between what Angoss calls the champion and the challenger. This experimental data can then be used to drive further modelling which is more ‘optimal’. It’s a cycle, and it is dependent on having a feedback loop in place, and being disciplined about testing and genuinely seeking the ‘true lift’ you achieve from the deployment. What maybe has changed are some of the metrics you can look at now, and the proliferation of metrics. There are so many levels of response – clicking, viewing, sharing, and the ultimate metric (buying). It can be quite overwhelming. Also some responders are more equal than others – those that are influential in a large network have more ‘social value’. And the volume of activity makes attribution challenging – what drove the sale, was it this thing or that thing?

Analytics and Employee Rewards: I Love Rewards Int...

Tablet Content and Google Analytics