For the taxonomy, words from articles are measured vs Classifications datasets (a.k.a. Bags of words)
To gauge how a media is perceived, we, this time, measure what is said vs Perceptions datasets
Matching and weighting are making the score of the document. This is how we know where an article belongs and its level of expertise/severity.
AI to keep datasets permanently relevant.
The perception we have on an event evolves over time. So should the classification of this article. To make it happen, we keep the datasets always up-to-date. As seen in last week post, the classification of an article evolves with time:
The evolution domino effect: Languages impact datasets, which impact taxonomy.
The way we talk about something evolves with time. In fact, the dialect of the tribe made of people concerned about something evolves over time. More specific, cooler, geekier, newer… whatever is the reason, we love changes, we love new ways to express ourselves.
Datasets updates. Permanently.
Datasets must be rebuilt and updated permanently to catch new words, new importance on words, up and down, and the disappearance of some.
Overlapping intelligence. On watch.
As datasets evolve they may make a classification overlap over another which will weaken the taxonomy.
We measure overlapping, permanently as well, and if one is detected the taxonomy evolves consequently.
Distrust in media is an issue everywhere. While trust in media in the US has stabilized around 40% (who trust), it was in the 70s-ish% in 1970. The situation is and is getting worse in Europe, as this Liberation article says :
“While there is also a downward trend of between 2% and 4% in most European countries, France is experiencing the largest drop in confidence. Above all, with only 24% of French people trusting the media, the country is 37th out of 38, just ahead of South Korea (22%). By way of comparison, the confidence rate is 47% in Germany or 40% in the United Kingdom.”
No trust in content, No trust in decisions made from it.
Can you imagine betting your future, the future of your business, on content you do not trust?
Can you imagine displaying your brand, the brand you’ve spent years building respect and trust on, on environment you don’t know, that do not fit your brand’s values?
Can you imagine having your PR and watch team listening media without understanding the profile of those media?
It is the motto of TrustedOut. The reason of our name: “If it’s not trusted in, it cannot be trusted out.”
Trusted Content should be like water or electricity: A Utility.
You need water. You open the faucet. You do not test the water. You trust it. You simply use it when you need it. Anytime. All the time.
You need electricity, You switch it on. You do not test the electrical. You simply use it when you need it. Anytime. All the time.
Imagine Trusted Content the same way.
You need trusted content. You open TrustedOut, define what you trust so you can trust the content you get. You simply use it when you need it. Anytime. All the time.
We now offer Unlimited access to TrustedOut, so access to content you trust is totally frictionless.
Define the content you trust for every segment of your business.
Hereafter is an example of an Enterprise organized by Industry. Also applies to any other type of business organization.
We won’t debate on the two articles under the scrutiny of the Singapore government, but rather focus on one thing very important for us:
Censorship must be and remain personal.
It is always dangerous to leave to someone what you can read and cannot.
CNN reports: “Government ministers can decide whether to order something deemed fake news to be taken down, or require a correction to be put up alongside it. They can also order companies such as Facebook (FB) and Google (GOOGL) — both of which opposed the bill — to block accounts or sites spreading false information.
The government can also prosecute individuals with fines of up to 50,000 Singapore dollars (about $36,000) and/or up to five years in prison. If the alleged falsehood is posted using “an inauthentic online account or controlled by a bot,” the potential fine rises to 100,000 Singapore dollars (around $73,000), and/or up to 10 years in prison.
Companies found guilty of spreading so-called fake news can face fines of up to 1 million Singapore dollars (roughly $735,000).”
Again, we, TrustedOut do not defend the spread of fake news or any offending content but we believe, for the most part, news can be seen as fake for some people and not fake for others, thus, censorship should be and remain Personal.
Get information from Traditional Media, have conversation on Social Media. Not the other way around.
In a previous post, we wrote:
“Misinformation and biases infect social media, both intentionally and accidentally. This highly recommended article from The Conversation exposes 3 types of bias identified by Indiana University. Hereafter are our takeaways… Continue reading
Trust, Media and Democracy
Related to this matter, we also wrote on the excellent Knight Foundation Report.
The Aspen Institute and the Knight Foundation recently released a report on a commission they organized about Trust, Media and Democracy. While coming from America, we believe most can apply wider.
If you don’t have the time for the length report, this medium page is very interesting. Here are our takeaways in the light of our previous posts, regrouped in 3 main categories:
Quotes using part of those speeches will, to be accurate, reuse keywords optimized for Search Engines, such as Google et al. Those keywords should prompt results where the political candidate and party will have better results, higher, in the 1st page.
Propaganda landing pages.
Those keywords, when searched in Google will be either those from a competitor where confusion can be introduced, but some will be usual in a context or out of date or simple fabricated. The more unique they are, the better SEO will operate. Rare, those keywords when searched, will return few, low inventory pages where the candidate crew will have much of the pages ready.
Talk the Google talk.
Better, words used in a speech will be seen as belonging to a specific wing. This way, the result is optimal:
Push the candidate
Push the point of the candidate with differentiation vs others
Push the party and augment bipartisanism
The cure: Watch who’s publishing.
Getting news from a search engine is very risky. SEO technics explained above and here in a previous post:
Let say you are looking for a media in “Eating and drinking”, in France, to associate your brand with.
1. Select potential media partners
France and French. Taxonomy is Eating and Drinking. We will select “Covered” for media covering this subject and Past month, for a more stable taxonomy than the last 7 days.
TrustedOut offers 40 media and 60 sources, 214 articles per day and 43,000 articles in the archive. Corpus creation looks like this:
Click on “Get” and we get Media and Sources.
A click on a media shows how the media is perceived for toxic contents and political orientations.
2. Diving on your selected two.
From the media list, you want more info on two: Le Figaro Gastronomie and Elle A Table.
A click on a source shows the trends in classifications [ 1 ], week vs month and month vs quarter, as well as the top classifications per period of time.
3. Get an instant report on a Media Source
For each source, click on the “Report” button, [ 2 ] in the screenshot above, and receive the report in a PDF format:
Caution: Report timestamp is very important as data are permanently updated.
4. Comparing 2 sources profiles
By comparing profiles, you can define what is best for your operations.
TrustedOut helps your business all the way, all the time.
Let say here, your company, SmartBizFurnitures, builds and commercializes high end, classy Business Furnitures. SmartBizFurnitures believes Lifestyle applies to Business environments.
Sources you trustto get reliable intelligence.
You must trust the content you use to trust any decision you make.
SmartBizFurnitures asks TrustedOut for all media sources covering:
People › Lifestyle › Decoration And Design And Architecture
Because they are looking for insights from their intelligence tool (Digimind, Netvibes, etc…), they decide to:
Solely select the classification your business is in: People › Lifestyle › Decoration And Design And Architecture
Pick “covered” to get a large but relevant number of sources for our analytic tools.
Go with the shortest period of taxonomy, the rolling “Past week”, because they want to be news-sensitive.
TrustedOut finds 481 sources. Click on “Get”…
Intelligence tools ready.
… download the CSV file of those 481 sources and import them in your semantic/social intelligence tool. Your tool will continue to listen to all new articles from those sources. Should you need articles abstracts from the past, ask us as we archive everything we’ve computed.
This download of sources is manual and thus your list of sources won’t be updated unless you do regular download and import to your intelligence tool. To avoid this and always analyze the freshest, most up to date sources, select “Connect”.
Media you trustto keep your brand safe.
Here you want the widest AND most relevant (read consistant) list of media for the safest whitelist you will feed your ad server with.
This time ask TrustedOut for the whole classification group: Lifestyle, but this time we want to get the largest but secure list of media who have published on Lifestyle in the past 7 days.
TrustedOut finds 3,717 media (and 10,037 sources)
Brand cannot be with “Politics”
SmartBizFurnitures does not want to be involved into “Politics”, so, those 7% should disappear from our Corpus map.
Let’s remove the whole classification “Politics”… with the 4th line: Media covering Politics in the past 7 days should be removed from the list:
Of course, we have less media: 936 instead of 3,717 but it’s the choice made by SmartBizFurnitures to keep their brand safe and consistant.
Same as with Sources. Here you will download the Media list as CSV but, as your brand must remain safe all over an ad campaign, you will want to “Connect” so the list of media = the Whitelist, will be up to date at all time.
Articles you trust to build on your PR efforts.
For our reading. we want the latest articles from the most relevant sources.
We will ask TrustedOut to give us our classification, as with Analytics, People › Lifestyle › Decoration And Design And Architecture, but since here, it’s not a machine analyzing but humans, likely a PR team or directly executives, we will select: Dedicated over the past 7 days.
100+ a day. 20,000+ archived.
On average, TrustedOut will provide 102 articles abstracts and can provide more than 20,000 from archives upon request.
A click on the right “Get” gives us:
Spot on for your PR efforts.
Click on the abstract article to read the article. Very interesting for SmartBizFurnitures. PR effort right here.
Watch tool ready.
Simple. Download: RSS.
Use this RSS as any RSS. Corporate pages, alerts, newsletters…
As of today and over the past 90 days, here are the top “covered” group classifications. This means the percentage of media in America covering the following group classification (like a countrywide newsstand shelf)
In bold are the differences above 5%. This means group classification that are more covered in the US vs France.
Top 3 Most topics covered are very different from France
US Media cover much more Entertainment & Leisure, Society and Education and, to a less degree, Sports, Lifestyle and Medecine & Health.
In bold are the differences above 5%. This means group classification that are more covered in France vs the US.
Greater coverage in Economy & Enterprise and Tech.
French Media cover much more in Economy & Enterprise and Tech which, both, are not even in the Top 10 US.
Largest Media Coverage differences.
As for above, we deliberately put a cliff at 5%.
Comparing Apple to Apple.
Tables above prove coverage between countries are very different, meaning feeding your intelligence tool and creating your whitelists based on random numbers of sources or media or articles will drive to unreliable, dangerous outcomes.