When you track flawed data, you risk informing the wrong decisions. What should you pay attention to during a Google Analytics audit?
Analytics | Jul 27
Victoria Lefevers on October 27, 2015 (last modified on February 17, 2016) • 6 minute read
A now classic example of this is the story about how Target’s marketing analytics department figured out how to predict pregnancy via data mining, and how ultimately they used that knowledge to target expectant and new mothers, which led to a rise in Target’s revenues from from $44 billion in 2002 to $67 billion in 2010.
Of course, that’s not the whole story. Target did manage to anger quite a few of their customers, most notably a father whose 16-year-old daughter received coupons for baby goods. Other customers reported being rather suspicious of getting a book of coupons where every single one was applicable to them. It doesn’t matter if the data is correct – people simply don’t like feeling like their every move is being tracked. It’s creepy. Going solely by the numbers can be a dangerous proposition – there are real people behind those numbers with varied experiences, perspectives and emotions. Numbers alone can’t give you the complete picture, but qualitative data can help fill in some of the gaps.
Take the Target example, for instance. Pregnancy is a sensitive topic for many reasons. Last time someone asked if I was pregnant, they almost got slapped. If Target were to send me coupons to that effect, they’d probably be getting a nasty email from me, regardless of if they were right. If Target’s marketing team had asked a handful of people, “In what situations would the topic of pregnancy be especially sensitive?”, they probably wouldn’t have wound up with a PR mess. Maybe I watch too much “Teen Mom,” but “underage pregnancy” is the first thing that comes to mind. A quick (completely unscientific) poll of my friends at dinner also brought up “IVF,” and “a history of miscarriages.” It doesn’t take a mind reader to figure out common sense issues with your customers – talking with just a few can provide you with a good gut check about how to approach delicate topics.
In short, qualitative inquiry accounts for the humanity of people, and can provide you with guardrails in which to unleash your quantitative genius. Admittedly, I am totally biased towards qualitative research (as evidenced by my habit of chatting with strangers, which my husband both loves and hates). The good news for those of you who aren’t big talkers is that you don’t have to do a ton of it to get great results. Here are some of my favorite ways to pair quantitative data with qualitative research:
Market research paired with user interviews help you gain a full perspective of where to play. Researching market size is a critical first step in establishing your total addressable market, but inevitably, people tend to make their TAM projections too narrow or way too wide. You can gut-check your market sizing by following up your market research with prospect interviews, which will yield a far more accurate understanding of whether your product is truly needed in the market. So make sure you include conversations with prospective users in any market sizing exercise. You’ll be happy you did when it comes time to start building your forecasts!
Win/Loss analysis, which is inherently both quantitative and qualitative, is a great way to figure out why customers make the choices they do. The best place to start is by simply emailing people who recently purchased (or churned) and asking an open-ended question about their experience. You can also try using a quick survey to gain initial customer feedback and open them up for a full conversation. Once you have that foot in the door, use their initial responses to probe deeper. Customers really do love to talk, so use this to your advantage and keep them talking until you have a very clear picture of why they made their decision. You should then pair some quantitative data with those interviews to really get a clear picture of what you’re doing right – and what you need to improve on. Some businesses use bookings to benchmark against win/loss, but don’t be afraid to use demand/lead gen statistics, share of voice, or another metric that is important to your organization. Pro tip: If you can, use a third party for interviews, particularly if you’re having trouble connecting with lost customers. Lost customers tend to be more responsive if they feel like they won’t be put on the spot by the company they rebuffed.
Primary competitive research paired with firmographic information on your customers (such as type of industry) will help suss out potential targets, forecast barriers to entry, and help you create a product strategy. I totally admit another bias here, since competitive strategy is sort of my raison d’etre, but I’m always amazed at how many young businesses ignore their competitors. If you want to play the game, you’ve got to know your opponent! Contact your competitors and get them to discuss the problems their products solve, paying particular attention to how they approach you as a potential customer. In doing so, you’ll be able to map competitors to specific audience segments. If you then pair that knowledge with information about your own customers, you can build a deeper understanding of your existing customers and your best opportunities. Are they healthcare offices? schools? Independent businesses? Where have you and your competitors penetrated? Where are barriers to entry low? Creating a deep understanding of your competitive landscape AND your target customers will ultimately guide your GTM strategy.
Customer conversations paired with surveys allow you to take a pulse of pretty much any aspect of the customer’s buying journey. Surveys are a fundamental tool for marketers. If you’ve built at least one, you probably intimately know how frustrating it is to realize (too late, of course) that a question which seemed clear to you was not at all clear to your respondents. Or, like many of us, you’ve wound up with 92% replying “other (please specify)” on a multiple choice question. To help avoid those scenarios, have a couple conversations with the group of people you’re surveying first. Listen to the language they use as they describe the things you’re interested in asking about. Practice a couple questions on them. Ask them what options they’d consider in response to your multiple choice based query. This will save you tons of time and avoid wasted responses.
In short, data is not useful without insight about the people it describes. The secret sauce to any successful campaign is to have both numbers and narrative at your disposal.
Analytics | Jul 27
Analytics | Jul 5
Analytics | Jun 29