There is such a large amount of information that can be gathered using social media about your customers. You have a pool of people – many of which who have personally invested in your brand already by following you. This investment usually means they like you enough to get regular updates and most likely want to see you succeed.
All of that aside, there are good ways to gather information and poor ways to gather information.
One thing I see a lot of is surveys. With tools like Survey Monkey and TwtPoll they are quick to set up and offer easy availability of data. These do-it-yourself solutions can be an excellent alternative to pricy research projects or firms for smaller initiatives. If set up appropriately that is.
You can completely sabotage yourself without even knowing it if your survey’s aren’t set up correctly. Remember, if you’re making your decisions based on poor data then you are making poor decisions. Set yourself up for success by designing your survey’s properly. Here are a few tips to get you moving in the right direction:
1) Determine Your Goals
You are probably starting with a question you want to get answered. Is this really the question you want to ask your customers? Likely not. What is the end result you are getting after? Design your project with this end result in mind. The goals will also determine what questions you include as well as who you ask, how and when.
2) Determine Your Sample
If you’re using your social channels you need to consider your sample and possible overlap. Are you planning on including your email list? If so, what does that group look like when you compare it to your Facebook . You’ll need to give a good deal of consideration to the size of sample you’re polling. As with nearly any data analysis, your results will have a great deal more meaning if you are pulling from a larger sample. Things like time and budget will likely influence this, but whatever size you do pull from you’ll need to consider this in your analysis afterwards.
3) Dealing with Bias
You’ll need to make a deliberate effort to eliminate as much of the bias from your survey as possible. It’s nearly impossible to eliminate all of it – especially in social channels given your likely respondents – but you must make ever attempt to limit it. Recognizing this bias exists can also help you take it in to account during analysis.
4) Determine Your Methodology
You’ll want to be sure that your methodology is aligned with the goals you have set. In addition you need to determine the best way to collect your data. Personal interviews, telephone/email/snail mail surveys and computer direct interviews are all examples of types of data collection methods. If you’re using social channels you’re far more likely to be interested in online questionnaires and surveys. Keep in mind that the nature of your research, if done completely on social channels alone is already biased to that subset of your customer base that is active on social media and computer/internet savvy.
5) Determine Your Questions
This is perhaps one of the most important pieces of your set-up. The questions you ask, and how you ask them will GREATLY influence the answer you get. Be sure you’re not asking loaded questions (those that are biased to generate a particular answer) or confusing questions. Determine what type of questions (multiple choice, rating, open-end or numeric open-end). There are considerations you will want to give for each type. I advise investing some serious energy in your questions and choose them deliberately and cautiously. You want to be sure that you are keeping your customers engaged so they complete the survey, unconfused (which leads to frustration), and also that the survey itself is not influencing their responses.
Much like you have a second pair of eyes read-through a guest blog post before you submit, you’ll also want to do a small test run of your survey. Whether it’s your friends over IM, office-mates or even spouses at home, get others to go through the survey and get feedback. They haven’t been elbow deep in it and will have a fresh perspective for you and are far more likely to catch errors or biases you may have missed.
Now it’s time to distribute your survey or questionnaire. Consider all of the different places you could push this and how that will influence the results. Facebook, Twitter and Email along with your forums and blog are great places to reach existing users. But if you’re trying to reach prospective users you may need to use a third party to help you get this survey out.
This piece of the puzzle will be nearly completely determined by the questions you asked. Some things you MUST consider during analysis regardless though are bias and sample size. Don’t take for gospel what may have been survey-designed outcome.
Some final considerations for you: keep the survey as short as possible while maintaining its integrity, consider your audience and what will most likely help you get clean and meaningful information from them, be careful to leave emotion out of question design, consider the privacy of your participants, and don’t forget to consider the participants motivation for taking part and the biases they may bring to the table.
Market research is a very complicated science. I tried to pull out only fundamentals to help those unfamiliar with research design. I admittedly am unable to cover everything here; rather I wanted to offer up some considerations to get you thinking more deliberately about your research methods.
Kristy Bolsinger is a Senior Associate at PwC in Seattle, WA. She has previously worked at Ant's Eye View (acquired by PwC in 2012), and RealNetworks (GameHouse). Prior to her time at RealNetworks, and Ant's Eye View - Kristy was working as a Social Media Marketing Consultant and completing her MBA at Willamette University. She maintains a social media blog and can also be found on Twitter, Facebook and LinkedIn.