Goodbye static, Hello dynamic!
May 5, 2010 | 3 Comments
Remember the good old days, the days of surveys? The good old days when you sent out 1000 surveys, waited two weeks, and were guaranteed to get 300 back but never more than 1000 back? That was nice, wasn’t it?
Social media research is a completely different story. We don’t send out surveys; we send out crawlers. We don’t send out questions; we bring back answers. We don’t wait to receive 100 completes; we cross our fingers that we have space to hold millions of records. And those are just the records from yesterday.
The interesting thing about SMR is the quantity of data. In fact, there are so much data on the internet that even speedy and greedy Google hasn’t managed to collect it all. Researchers, on the other hand, are far more careful and picky about collecting data. We still get excited by the minute as data come in, and it does come every minute, but we carefully observe it as it comes in.
We notice that some of the data are fresh out of the oven, created just minutes ago, while other data are less fresh perhaps created hours or days ago. And then, there are always the data that were created weeks or months ago. This is the data that was written with just as much passion as other opinions found long ago, but had been hiding behind poorly tagged blogs and rarely indexed folders. It was there all along, valid and important opinions desperately trying to be heard and appreciated, but only found now.
Data gathering from the internet is full of highs and lows. Some days, data pours in by the truckload while other days, it comes in more slowly. Some days, data is as fresh as fresh can be while other days, it is discovered from long ago like ancient treasures of Egypt. No matter when the opinion was offered though, researchers know that it’s important to count every opinion.
That is the interesting thing. You used to know that the 300 surveys in your hand would still be 300 surveys tomorrow. Social media has turned that traditional data rule on its head. Three hundred verbatims today might be 350 tomorrow and 600 next week. And, that’s ok. It’s better to have collected a larger proportion of the population than to have unknowingly excluded subsamples that might completely change the results. It may not be how we’re used to collecting data, but researchers too must change with the times.
Live data? It’s a wonderful thing.
Category conversition | Tags: dynamic data,sentiment analysis,social media research,static data
Social Networks : Share
3 Comments so far
by Brandon Watts
On May 5, 2010 at 1:11 pm
Great post. SM research is definitely an exciting frontier for research geeks! One Question: Are there concerns with how representative the SM sample populations are? While SM undoubtedly provides a larger overall sample size than an old school survey (n=300) and an ongoing stream of data, it seems difficult to imagine that the populations (active SM “commenters”/participants) are representative. It’s not a question of whether SM research is valuable or not – clearly it is. It’s a question of how far one can extend the resulting interpretations.
by Conversition Team
On May 5, 2010 at 1:14 pm
As with any marketing research these days, representativity is a difficult topic. Most surveys can’t claim probability sampling and therefore can’t claim true representativity and neither can social media research. In the end, it comes down to researchers having the skills to properly interpret data given all the variables associated with a particular study.
And, we love being called research geeks!
by Brandon Watts
On May 5, 2010 at 1:56 pm
Very true. Representativity is indeed a challenge in all modes of market research, and in many cases the cost/feasibility of the goal far exceeds the value. Obviously, it depends on the objective at hand, but the reality is that clients need reliable information to inform key business decisions (and usually/always need it FAST). Also agree that research geek is a compliment!