We use human analysts to understand online discussion

Understanding what people say online is essential. If you just count mentions or use high-level automated analysis you will never get to the nuances of buzz.

At the core of the WaveMetrix methodology lies the fact that we use human analysts to read every relevant comment made online. This is obviously time consuming, but it is also what delivers value to our clients because it enables us to move from monitoring social media to really understanding it.

Mother tongue analysts only

All WaveMetrix analysts are mother tongue speakers of the languages they work in.

We only use mother tongue speakers because language online is typically very informal and contains a lot of slang. To properly understand it, an analyst needs to be able to not just read the words, but also appreciate the cultural context.

Advantages of human analysts

Why do we spend so much time (and money) having analysts read every post? We think there are five advantages to having human analysts:

The five advantages of human analysts

• 1. Topic. Human analysts genuinely understand what is being said and can therefore code all discussion properly. They can take consumer comments of any length and break them down into constituent parts. By classifying buzz in this way human analysts formulate data which is very granular, meaning it consists of many precise and specific elements, and is therefore incredibly powerful
• 2. Sentiment. WaveMetrix analysts code sentiment according to action on a normative scale of +2/-2 (see section on sentiment). This is done for each opinion and consequently provides data on consumer sentiment which is very accurate
• 3. Complex categories. Analysts can identify more than just topics and sentiment. For instance, they can isolate recommendations, whether explicit or implicit. Understanding recommendations is extremely valuable to our clients because they are often strong drivers of sales. See sections on topic and advocacy
• 4. Alerts. Human analysts deployed on a continuous tracking study can very accurately identify important posts. This means they can trigger alerts which allow clients to respond in a timely fashion
• 5. Comprehensiveness. Sometimes it can be difficult to capture all buzz because many conversations do not reference the brand or product they involve. One poster may reply to another, referring to a brand as “it” or “they”. Posters may make references such as “David’s car” which would be obvious in context, but difficult to find if relying solely on search technology. Human analysts are much better at ensuring comprehensiveness in these circumstances than technology. This is one of the reasons why we use a combination of search technology and human analysts to identify buzz. Another key reason is that analysts can access sites where robot crawlers are banned, such as many review sites 

Managing human analysts

We regularly run projects which span up to 20 countries (and occasionally more). Ensuring that all analysts work effectively, that they understand the client classification scheme and code findings consistently is not a straightforward task.

At WaveMetrix we are very proud of our ability to manage mother tongue analysts across the world. We think it is the hardest thing to do well in social media monitoring, we invest a lot of effort in doing it to an exceptional standard and it is one our core competences.

The WaveMetrix system for managing human analysts

• 1. Dedicated research executives. All WaveMetrix projects are run by a client-facing research executive. We do not split client contact and project execution, because we think it is essential that client needs are communicated as directly as possible to our analysts. We also think clients generally prefer to talk to the person who will actually execute their work
• 2. Bespoke search and workflow environment. Our search and workflow environment enables analysts to work effectively in classifying online discussion. Driven by buzz search tools, the workflow tool allocates work among analysts and establishes an audit trail to ensure buzz is captured comprehensively and accurately. Our search and workflow environment draws extensively on our proprietary databases of online buzz. This ensures that our learning from the thousands of projects we have executed in the past is reflected in managing analysts to find online buzz on current ones
• 3. Customised classification software. We build dedicated classification software for each project we execute. This ensures coding of online buzz is accurate and in accordance with the classification scheme agreed with the client, whatever the country we’re working in
• 4. Quality assurance team and processes. Every project we run goes through a strict quality assurance process. Through multiple checks involving our in-house QA team and the dedicated research executive, we ensure data is accurate and comprehensive