Sentiment analysis · Total exposure tracking
i‑sieve crawls the web (news, blogs, forums, social, video and audio) and finds exactly what people say about your company, your campaigns and your competitors. Then we hand you the data.
How do people react to your advertising?
We can tell you.
What do people really think about your company?
We can tell you.
How would a stranger describe what you do?
We can tell you that too.
What we do
The buzz‑phrase is sentiment analysis. Here’s what that actually means.
We identify relevant comments made by news media and by people in blogs, micro‑blogs, online video (in the stream itself and in the comments beneath it), discussion forums and public social networks. Sophisticated techniques extract what’s relevant to your question and classify it as positive, negative, neutral or balanced. Everything irrelevant is discarded.
Live classification · campaign: “XYZ ad”
“I love the music on the new XYZ ad. Anyone know what it is?”
relevantpositive“The new XYZ ad is SO SEXIST!”
relevantnegative“Hey guys, have u seen the chick in the new XYZ ad? She so looks like Sally!”
relevantneutral“Products on sale in the new store include ABC, DEF and XYZ.”
irrelevant
The last one mentions the product but says nothing about the campaign, so it never reaches your report.
Total exposure tracking
Every channel your brand appears in. Measured.
Buzzsense normalises simple mention counts into meaningful insight: sentiment polarity, the paths and patterns of content spreading, and the size of the audience reached.
Online text
Social and professional media, side by side.
We measure and analyse social media alongside professional online media such as news portals, tracking brands, products, campaigns and public events by the impact they have on the online community.
Delivers · Sentiment polarity, plus the drivers behind positive and negative comment.
Video
Every second your logo is on screen.
We analyse digital footage to identify brand exposure via sponsorship (sideboards, overlays and logos in sports coverage) or via product placement in TV and film.
Delivers · Exposure in seconds (time‑in, time‑out) and screen location, with variable occlusion and logo‑size thresholds. Footage capture available on request.
Audio
What is said, and how it is said.
We analyse digital audio signals to detect references to sponsors, people and brands, down to the second.
Delivers · Keyword and key‑phrase mentions with contextual sentiment (via speech‑to‑text) and emotion (via voice analysis). Stream capture available on request.
Print media
Publications, digitised and measured.
We measure exposure in publications available online in their print form, free or by subscription, and can digitise print media for you.
Delivers · Sentiment and its drivers, publication metadata (page, journalist) and photo exposure: the size and location of your logo in every image.
How it works
A three‑step process. The web is our database.
Defining the search
You define the “category” in a standard brief. It can be wide (compact cars, horror movies) or narrow: the new Mini Cooper, a single film title. We run early probes on the web to augment the description, double‑check the findings with you, then our machine‑learning algorithms build an ontology for the category from both the brief and the content discovered.
Collecting & analysing
Our proprietary focused crawlers use the ontology to separate mentions from opinions across roughly half a million online resources. Content is processed, classified and analysed: we calculate clusters, measure sentiment polarity and identify sentiment drivers: which aspect of the brand, product or person drives positive or negative opinion. Results are normalised by the buzzfactor.
Reporting
An online dashboard with key volume and sentiment metrics, with the actual data underneath it. Every comment, its URL, its buzzfactor. Alongside the data we provide our own analysis and conclusions and, where requested, graphical representations of public sentiment.
We don’t just give you a chart. We give you the data.
One of i‑sieve’s unique features: we deliver the exact data our classifiers used, usually as a spreadsheet. The comment itself, its URL, and a measure of its impact we call the Buzz Factor. Captions for text, annotated proof images for video.
You’re free to analyse it yourself, check our findings and see every comment in its original context.
~500k
resources scanned per project
2–7 days
to train a new category
Daily
reporting once trained
Who we work for
Anyone with a message to convey, a brand to promote, a policy to implement, or a point of view to popularise.
Case study · 26 August – 2 September 2009
London Climate Camp
A protest organised via social media, with the location kept secret until the last minute. We tracked reaction to the companies named by participants and observers.
2,800
Relevant references
311
Named specific companies
1,700
Relevant tweets
All web references
Shell · Twitter
“Live from Climate Camp: four activists have climbed onto Shell entrance and removed ‘S’ on sign so it reads hell centre”
twitter.com/standardnews
Start here
Know what they’re really saying.
Send us one question about your brand, campaign or issue. We’ll come back with how we’d train the system for it, when the first data lands, and what it costs to keep listening.
- No obligation
- Any language
- You keep the data