Dave asks, "How do you interpret the "why", i.e. if you see data that says actions were taken (or not) how do you determine WHY those actions were taken so you can replicate or avoid those specific marketing tactics moving forward?"
No analytics tool focused on gathering "what" data - meaning, what happened - is going to give you "why" data - why customers made the choices they made. In this video, we walk through three data sources to get "why" data, and places you'll need to go to get it.
Can't see anything? Watch it on YouTube here.
Listen to the audio here:
- Got a question for You Ask, I'll Answer? Submit it here!
- Subscribe to my weekly newsletter for more useful marketing tips.
- Find older episodes of You Ask, I Answer on my YouTube channel.
- Need help with your company's data and analytics? Let me know!
- Join my free Slack group for marketers interested in analytics!
What follows is an AI-generated transcript. The transcript may contain errors and is not a substitute for watching the video.
In today's episode, Dave asks, How do you interpret the why if you see data that says actions were taken or not? How do you determine why those actions were taken? So you can replicate or avoid those specific marketing tactics moving forward?
Good question. The answer is from the existing marketing data you have in your analytics tool, the answers, you probably can't. And the reason why is that analytics tools like Google Analytics, for example, are really only collecting the what what happened, not the why. And for the most part, you can't extract a whole lot of meaning from that, with one or two exceptions, which we'll talk about in order to extract the why you have to get inside the customers head and that means asking them
questions and and looking through and researching the answers. So let's go through a few different methods of doing this what's called qualitative data analysis. The first and and easiest to implement is on all your intake points, are you asking customers how they heard about you? Is there a box where someone can type in I watched a webinar with you, for example, if you're not doing intake, assessment, intake data collection, you're missing an easy opportunity
to get some of that data and to understand, okay, this, this is useful stuff, be very careful about drop downs, I don't like drop downs, because what tends to happen is that people choose the easiest thing and not necessarily, you know, the thing that is most prominent in their memory. So that unaided recall, how did I hear about this company, I find to be very, very useful.
The second is to mind the data you are already collecting
you are collecting, I would assume customer service data, call center data, CRM data, and so on and so forth, you're getting a lot of text input
from customers, your customer service inbox, maybe even key executives have data that they're collecting. Now, they they get emails from, from industry, peers, and so on, and so forth.
And so there may be options there, you want to gather all this data, pool it together, and then use text mining tools, machine learning tools
to extract out hey, these are the common words and phrases over time I keep
appearing are disappearing and being able to quantify that and then match that against
calendar, your editorial content, hey, we published a great number of pieces of out about the new being a coffee that we released
n. And sure enough, two months later, customers were emailing us about our our new coffee. So that would be an example of using some of that existing data. The one
the one thing I'd be very careful of there is that you if you're going to use social media data, be very cautious with it. Because it is extremely messy. There are significant problems with bots and automation, meaning that you may not be analyzing data coming from real people and,
and social media networks themselves are biased. Each network has its own bias for things like age, and gender, and demographic. But then also there's behavioral biases to generally speaking, people don't participate on social media about stuff about a brand unless either it was a really powerful activation
or they really happy or they're really unhappy. And if you're trying to do something, for example, like our FM analysis, where you're trying to take
okay customers and turn them into great customers, get them to buy more, you're going to have a big gap in your data, because the satisfied but not thrilled customers aren't going to bother posting their it's simply not interested. And so be very difficult to identify those people. So that's mining the data you have. The next thing, of course, is getting more data getting more qualitative
data. And this comes from things like focus groups for Are you writing focus groups
that are sitting people down and having a chat for a couple hours about hey, this is what they are looking for in a next brewed coffee?
Are you doing stuff like that, where you can get information from them that is qualitative in nature, that you can then say, Okay, these are, this is what our customers really care about. So let's make sure that we do less of this and more of that
customer advisory boards. I think probably one of the most underrated tactics for getting qualitative data about your brand and your marketing. When you have customers who are already paying you money, you can ask them questions about your marketing and say, hey, did this resonate with you? We ran this campaign did even see the campaign
How would you have done it differently if we want to reach more people like you because your best customers, how do we reach more people like you. So customer advisory boards and think are
powerful. And then of course, if you want to kick it up a doing formal market research with a credible market research firm is certainly a way to go for going to spend $10 million on your next you know, massive media by Shouldn't you spend 100 grand on making sure that the message and the concept and everything correct this the firm I recommend there if you're going to do something like that would be a company like Edison research Tom Webster shop, but they are top shelf and and get you credible,
the when it comes to the qualitative data, it really is asking people it is getting insights and data already happening, collecting new data on top of that, that's how you get those qualitative data insights. And then what you have to do is, then start to quantify them. So if you have a focus group where you have some mind data, qualitative data
is not quantitative data, which means that you can't just slap a you know, say 55,000, the seven customers we asked in our focus group want this,
but they may still be biased. So you still want to run a larger scale quantitative survey emails to your customers or surveys, through survey panels or market research firms to validate your qualitative findings. It's an ongoing process in order to make sure that you're doing and you're getting the right data from your customers and from your audience that will lead you to believe to do
get correct answers. So good question. It's it's a complicated questions, I'll probably a lot more complicated than you were hoping for
the one the one area where you can get good insights about what's working from your quantitative data is around attribution analysis. If you have the data and you have the software to pull it together,
if you have the data means that you are a primarily online business, like if you're going to analyze data with Google Analytics in your primarily online business.
And then there's very little interference from outside sources that are not online. And you can use attribution analysis, either built in or run more sophisticated attribution analysis to figure out what's working. If you have all the significant components of offline, you have to gather all the data from all the offline sources, put it in a massive spreadsheet, along with all of your online data,
and then run
again, very sophisticated machine learning algorithms, one particular called gradient boosting that will help you understand what combination of activities and variables and things leads to the outcome you want. It is doable, it is possible but it requires a lot of heavy lifting to get it to get that data all pulled together. We're in the middle of a project right now for a customer and it is taking, you know, weeks to to build the models and test them and things like that. So it's not an easy thing to
but it is doable if you want to so good question, Dave. As always, please subscribe to the YouTube channel and the newsletter
I'll talk to you soon.
You might also enjoy:
- You Ask, I Answer: Quantifying Hallway Conversations?
- Almost Timely: The 2020 Essays
- You Ask, I Answer: Microsoft Clarity vs. Google Analytics?
- You Ask, I Answer: Creating Content for Search Engines?
- How To Start Your Public Speaking Career
Want to read more like this from Christopher Penn? Get updates here:
Get your copy of AI For Marketers