See It Coming.

We do macro-listening,
a new style of content analytics.

Using search, network science, and social theory,
macro-listening tracks changes in shared mental models
of audiences—not actions and attitudes of individuals.


See It Coming.

Big insights and big data are not the same thing.

Collecting ever-more and finer-grained individual-level data
can actually make things worse, not better.

This is because crowds present critical strategic contingencies
that do not reduce to aggregations of individual behavior or cognition.

See It Coming.

Use macro-level data to address big picture questions:

When do nascent markets or movements become real?
When do fomenting movements demand attention?
When do new realities become threats, not just opportunities?
When is it too late to enter a new market? Too risky not to?

Spot emergence of a new market before others do.
Be confident about the core elements of a new market concept.
Direct investment and investor communication to make your vision pay.

MemeStat is for studying changes in meaning

But why study changing meaning? Observing changes in the meaning of a meme from text supports inferences about changes in related attitudes or behaviors. Of course, understanding social change as it is happening is never as clear as with the benefit of hindsight's perspective, but the ability to see changing meaning in current and very recent public discourse provides at least some help for the job of understanding or even trying to anticipate important social changes such as the following:

  • Innovations that create products categories that become new markets
  • Business strategies and related capabilities that transform industries
  • Creative work that alters the landscape of literary or artistic genres
  • Scientific breakthroughs seen as establishing new fields
  • New ideas that anchor influential social movements or political issues
  • Beliefs that inspire and mobilize people to do great good, or grave ill

MemeStat's Main Functions are that it enables you to:

  • Mine your collection of news stories or blog posts (a "corpus") to produce dynamic semantic networks
  • Analyze networks for mathematical properties relevant to idea currency (see Kennedy, Lo and Lounsbury 2010)
  • Visualize networks to see relationships and how they change over time
  • Browse semantic networks to get to stories by clicking on visualizations or summary charts
  • Export longitudinal data sets for custom statistical analyses

How MemeStat works

To use MemeStat, you need ...

  • A corpus of news stories, press releases or blog posts from any combination of Lexis/Nexis downloads (html format), Factiva downloads (xml format), or URLs for blogs with RSS feeds.
  • An ontology that relates terms for the idea you are studying to those potentially relevant to its meaning—attributes, instances, synonyms, etc.
  • Your team! MemeStat allows you to build and join multiple overlapping teams to share work and results like a social media site for research.

What it does. MemeStat can analyze your corpus for two kinds of data about patterns of word usage over time.

  1. a hit-count: every mention of every term in a list of terms you want to track; presented in either a table or histogram format
  2. an association: every co-mention of every possible pair of terms from two lists of terms you want to associate;
    presented in a table, histogram or graph visualization format.

What you'll see and get. MemeStat allows you to explore a hit-count or association at three levels:

  1. A View (a table, histogram or network) of results by period (move forward and back) in which clicking on elements populates ...
  2. A list of Stories associated with the selected View item(s) in which clicking on stories populates...
  3. A Text viewer with the selected story with mentions or co-mentions highlighted.

Technologies that MemeStat Builds On

The heart of MemeStat is an association engine, a search utility designed to do multiple searches that show either how mentions of a set of words change over time (think of it as 'count this'), or how co-mentions of two sets of words change over time (think of it as 'associate this and that'). The association behind MemeStat is called AE.


  • From AI, the use of semantic networks (graphs) to model meaning
  • From sociology, the idea that structure reflects meaning, and vice versa

Open Source Libraries used

  • Lucene is used to build AE, MemeStat's association engine
  • MySQL is used to store and retrieve all data
  • Protovis is used for visualization of graph data. (We will soon switch to its successor D3 or Gephi.)
  • UCI's JUNG framework is used for social network analysis


A market takes off. A movement catches on. A bubble bursts.

These often seem sudden, but in hindsight, overlooked signs appear.

Macro-listening helps you see these things coming by
tracking shifts in audience understanding
of a concept, market or movement.


It works because language and social structure co-evolve.

This means that public discourse contains data about
coming shifts in fundamental social structures.

This means you can see changes coming in
markets, movements, fields, genres and types of organizations.


For whatever you want to see change coming to, we do the following:

1. Assemble a collection of relevant text—including history.
2. Run "associations" that captures shifting linkages among concept elements.
3. Analyze associations for emergence, change, stability.
4. Track and update to watch continued changes.

Example Projects

Here are a few of the projects we've done:

  • Shifting demand themes in consumer packaged goods categories
  • Emergence and changing meaning of "green" business
  • Emergence of smart buildings
  • Changes in concept cars by categories and features
  • Emergence of Tea Party versus Occupy
  • Email analysis of potential liability for common employment risks


Mark Kennedy, Chairman. Mark, an academic at Imperial College London, founded Textonics to develop new content analytics for both practical applications and research. PhD, MBA Northwestern; AB Stanford.

Richard Jones, CEO. Richard leads Textonics from experience gained as CEO of MediCorp and a consultancy he founded for innovators.
MBA Northwestern; BA Olivet Nazarene.

Jim Kennedy, Senior Advisor. Jim brings 40 years of experience leading change, strategy and software projects at Bell Labs, Booz & Co. and CSC.
PhD, Stanford; BS, Penn State.

For More

Fill out this form to learn more or request access for your research.

 First Name

 Last Name



 Confirm Password