Author: Wei Li, Chief Scientist
Automated surveys, using social media as sources, will eventually largely replace manual surveys. This is an inevitable outcome because social media has become the main outlet of public opinion, and technology continues to offer deeper analytical options for this readily available social data.
Automated surveys, or auto-polls, use computers to collect public opinions and sentiments on a topic. By parsing social media big data and mining salient information (facts, evaluations and emotions), we can understand social users’ views about any topic, and use that understanding to inform key business decisions.
This Artificial Intelligence already offers numerous proof points in the business world (e.g., our customer insight products), putting a new spin on a tried and true concept.
Why Polls Matter
Polls have provided quantitative information to guide decision-making in government, businesses, and the general public for many years – like during presidential elections, to cite a prime example.
Polls are conducted periodically throughout the campaign to inform voters, as well as the presidential candidates, about public perception of the candidates. This allows voters to make an educated choice, and the campaign teams to adjust policies and strategies to enhance their candidates’ public images.
In the same way, feedback collected from customer surveys can help enterprises detect issues and address them – during a product launch, for example.
However, this type of feedback is limited to singular events, whereas social auto-polling offers constant feedback that is also much faster, more comprehensive, and less costly – all at a larger scale.
Compared with traditional manual questionnaires or polls, auto-poll has some key advantages:
Real time. It often takes days or even weeks to complete a comprehensive survey manually – between designing/distributing questionnaires, conducting telephone or street interviews, and collecting and summarizing the results.
But auto-polls are instant, because they aren’t really polls in the traditional sense at all; they’re conversation starters – you get results as soon as you enter your topic.
As long as people are discussing your topic, social insights can be mined as easily as using a search engine, with the same response time, but yielding much more accurate results – because NetBase’s deep parser reads social media day and night to feed our storage, just as a search engine indexes the Internet in their storage.
This immediacy gives businesses a desired edge in dealing with consumer complaints, understanding consumer mindset/potential trends, and calculating best marketing strategies to employ based on consumer preferences.
Low cost. With manual surveys, businesses must weigh the cost of larger scale surveys (which reduce margin of error and are more reliable and convincing) against budget-enforced compromises in sample size.
Auto-poll affords the ability to analyze a variety of customers and topics for a fraction of the cost of a traditional poll (often with a sample size of millions of data points vs. several thousand data points – way beyond the reach of most traditional polls). So it’s not just that the cost is lower, it’s that the results are better as well.
Objectivity. Questionnaires designed for traditional polls or surveys may intentionally or unintentionally introduce subjective bias or implied suggestions.
Because auto-poll is bottom-up data analysis and mining – based on people’s natural comments on topics, not responses to specifically-designed questions – it is more objective by nature.
Traditional poll incentives can also introduce a possible bias via customers who answer simply to gain rewards, compromising the integrity of the poll’s results. Auto-poll happens without prompting, producing the most authentic feedback businesses can ask for.
Multi-topic comparison. This is particularly important because polling data means nothing if not compared against the data of industry competitors.
But to acquire a full brand picture compared to all leading brands in the same industry using manual surveys would require more time and money than is feasible – even though that information is crucial to gaining a competitive edge.
With data mining systems using auto-poll, surveying multiple brands or products is as easy as surveying just one.
This is how NetBase’s BPI (Brand Passion Index) feature works – instantly surveying multiple brands in one industry and comparing them across three dimensions: buzz (size of the bubble), popularity (up or down in the graph), and passion intensity (right or left in the graph: the more right, the more intense).
When brands understand the depth of feeling consumers have for them compared against their competitors, they can use that information to approach social conversations in a new way to attract their desired audience.
The best part is this information is available instantly – and with mobile-web and social media in everyone’s hands as we enter a new age of big data, being able to mine the public opinions and sentiments buried within this volume of data in real-time is paramount.
Luckily this technology – mature, large-scale, multi-lingual text mining systems that parse and read big data around the world – is already here. And as auto-poll becomes more mainstream, even bigger and better technologies will be just around the corner.
This blog first appeared here.
About the Author: Dr. Wei Li has been the Chief Scientist since 2006. At NetBase, Dr. Li has been focusing on leading the development of information extraction and sentiment analysis technology to power technology searches as well as social media analysis of brands.