One of the fundamental problems we have in our marketing efforts today is the difficulty in handling and interpreting the large amount of data we gather and have access to. A question that might center this problem more is how do we filter the unimportant from the vital data?
Metrics can add powerful insights into our marketing campaigns but they don’t solve our problems out of the box. In fact, they could create more problems for us if we don’t know how to tame them to gain insight into what they’re telling us.
What to do with the data?
Past generations of marketers who didn’t have ready access to all this information had to formulate slogans and jingles to get into their customers’ heads. They relied on their creative energy more, hoping that they had their fingers on the cultural pulse and could persuade people through their creative appeals to buy their products.
Of course these previous generations also had surveys and focus groups, but they didn’t have near the extent of information that we have in applications like analytics. They didn’t have the incredible research tool of search engines to wade through a daunting supply of insights and information on everything from specific information on industries and segments of the population, to customer behavior, insights from neuropsychology, etc. They also didn’t have the free willing disclosure of potential customer’s through forums, reviews and social media.
We’re truly living in interesting times.
The solution to winnowing the data chaff: Research
It’s ironic then in digital marketing that the solution to this problem of an overabundance of data can only be solved by more focused research. Only by researching in an organized way can we find insights into the huge amount of data that we have.
There’s few principles that we might want to think about when we start trying to figure out what to do with our data:
- We need to be problem-focused. The data that becomes important is that which can help us with our problems.
- We need to be open to have our views to be challenged in light of the research. We should be conscious of our biased perspectives as much as possible and be willing to give them up in lieu of what we discover.
- Last, we should be solution-driven. We need to ensure that the data we are trying to uncover is actionable. It shouldn’t be the data that drives, us but it should be our quest to better conversions and profits on our site.
CXL has developed a model of conversion optimization research called ResearchXL that maintains an awareness of these principles throughout. ResearchXL allows us to focus on the important aspects of our data in order to find problems that need solving on our sites and even helps us discover some potential solutions to these problems.
There are six steps in the ResearchXL model. They are as follows:
Step 1: Heuristic Analysis
The first step is to always go directly to the main source that we are trying to refine, not just our websites but all of our funnel process, including the sources of our traffic: ppc, facebook ads, social, etc.
What we want to do is go through each ad and web page and experience the entirety of our funnel through an empathetic, experiential lens. Peep Laja gives four factors that should inform our heuristic analysis.
We should assess our digital marketing project for the following:
- Clarity: Is it clear what the purpose of the site is, what it is offering, and what are its offering’s features and benefits? Does it have a clear value proposition?
- Friction: Are there things blocking the user from getting through the funnel? Are there unnecessary information on forms, for example, that might be preventing conversion?
- Anxiety: Is there a clear message match between ad copy and the website to ensure trustworthiness? Does the site answer any concerns that the potential customer might have about the product/service that the site offers?
- Distraction: Do all web pages have one overarching goal? Is there unnecessary copy or images that do not serve to help the user reach that goal?
2. Technical Analysis
The next stage of research serves to bring any technical problems to light. There are many different companies that run these tests for a fee and even some open source ones that work through collaboration.
Here are a few things that we need to analyze:
- Cross Browser Testing : We need to look at all commonly used browsers to ensure compatibility with the site and flag any problems that might occur because of incompatible code.
- Cross Device Testing: We also need to look at different devices such as mobile, tablet, and desktop and test various screen resolutions.
- Site speed analysis: We need to use tools like Google Pagespeed and Pingdom to assess whether our pages are loading efficiently or not. These tools not only diagnose many problems for us but they also offer solutions to these problems.
3. Digital Analytics
Google Analytics coupled with Google Tag Manager is important for any quantitative research that we have to do for our site. Analytics can give us clues into where people are dropping off on a site, they also help us understand what types of behaviors correlate with more purchases, and they can help us assess where friction might slow potential customers down from fulfilling page goals.
One of the things that we need to do though if Google Analytics is going to aid us in discovering these things is to assess whether Analytics is set up correctly to give us this information. According to Peep Laja because so many businesses have problematic, incorrectly setup Analytics accounts one of the very first thing we should do before getting into our CRO project is to health check the setup.
This involves basically addressing any issues that would prevent us from gathering insight from the website, including problems that might be affecting the data coming in.
4. Qualitative Research
Qualitative research involves running and cataloging surveys, polls, reviews, case studies. All of these help to refine our site, both meeting the needs of our customers and discovering how we can better position our product or service to make it more desirable.
We can use onsite polls, for example, in several ways to discover what friction might exist on our site. It’s important to run only one poll at a time on a site, so as not to create additional friction. But good places to put polls, include anywhere you expect friction to exist on the site: on the product page, the cart page, or anywhere there might be a specific concern.
We should also survey new opt-ins or purchasers within no longer than a few days after their conversion. Surveying during this time allows us to get reliable feedback. Eight to ten questions mapping out their journey from problem seeker to problem-solver could give great insight into how we can improve our buyer’s journey via our digital marketing program.
5. User Testing
User testing involves recruiting test users that represent your actual target audience to use your website.
Giving the users a series of tasks usually helps ferret out places where people might get stuck or frustrated on the site. Typically these users are given a general task like pick out something nice for a family member, and then a very specific task like find a pair of red, size 32 pants that cost less than $60 and order two of them.
Observation is key here. You want to see what they do on the site and if they fail to meet any objectives or they get bogged down by anything. Users might maintain that the site was very easy to navigate, while their actual actions on a site might tell a different story.
6. Mouse Tracking Analysis
Mouse tracking analysis tracks mouse movements on a page. Hover maps are not very useful, but other tracking like click maps that show where people are clicking on the site, heatmaps are based on an algorithms best guess as to where a user is looking given the movements of their mouse, and scroll maps show how far a user is scrolling.
Another valuable tool we can use as conversion optimizers are session video replays. These are videos of people working the way through the site. and working through parts of the site that many times need optimized like forms.
What to do After the Research
All of this research should yield insights into our data that we didn’t have before. Ideally, we’ll have discovered many problems that we can now set to solving.
Once we’ve found these problems we can develop a system to rate them and begin fixing the easiest, least expensive, ones that have the most impact on our profits first. Gathering data with this problem-focused, solution-driven research mindset keeps us from just collecting data for data’s sake and helps us to focus on how to get better and faster conversions on our site.