Do you have an analytics strategy? Analytics can help inform every web project, but only if you know what to look for. There is a huge difference between “analytics” and an “analyst’s report”. Far too many times they are used synonymously but the truth is, data ≠ analytics.
Data needs to be interpreted in order to become analytics and in order to do that, you need an analyst with some good insights. I have seen far too many reports that have interesting data, but there really isn’t any strategy or actionable insight that comes from the data. Data can be used for so much more than just determining if you are up or down over yesterday, last week, last month, last year… A component of analytics covers those details but data by itself does not tell the whole story. A good analyst will know what story needs to be told, they will figure out what information needs to be pulled and how to compile an accurate depiction of what is happening. Most of the time, there will be holes in the data, but a good analyst will look elsewhere to fill in the holes and come up with a conclusion based off of the data that they are able to pull. A good analyst will be able to collate data from multiple sources to support their conclusion, but a great analyst will adjust their conclusion when the data is pointing in a different direction. Analysts don’t always have to be right. The great ones will lead you in the right direction but come up with alternate recommendations if their initial thoughts are proven wrong by the data. Analysts will work with the business to understand what what Key Performance Indicators (KPIs) should be defined and then determine the best way to collect, parse or consolidate data to tell an accurate story of what is happening and how the business might be able to make adjustments to move the metrics in a favorable direction. You should include your analysts when you are in the planning phases of a project so they can define the KPIs and make sure the metrics that need to be measured are tagged.
The key to improving any web implementation is to make sure you have a good analytics strategy but even more so, a great analyst. Make sure your Calls to Action (CTAs) and KPIs are defined throughout the site. Identify the different paths that users can flow through the site and make sure you know what the KPIs on each of the major steps of the funnel are. In order to do this you will have to identify all of the functions that the website is serving for your organization; does it do web sales, marketing, establishing brand? Does it assist in sales and help recruit talent? Does it do all of this and more? Make sure that you define these functions and then define conversion funnels for each of these end goals. Different steps in the conversion funnels will have different purpose and different KPIs. Gary Angel has written a great white paper on this topic called “Functionalism”.
If you look at a typical eCommerce site you may see a sales funnel that looks like this:
Homepage>Category Listing Page>Sub-category Page>Product Detail Page>Shopping Cart>Shipping/Billing info page>Confirmation Page.
In the simplest form of a sales funnel that particular path would have 6 points at which a person can get distracted or leave the site. And that is the simplest form not including going back and forth between categories, adding product to cart, removing product from cart ect. In this 7 step sales funnel your end goals is a sale. That would make the conversion for this funnel a product sale. If you want to improve the sales in this funnel you can’t make a change to the second step and expect to see if your change was an improvement by looking at the sales results over the next few weeks. There are far too many variables at play within each step. Let’s say you are selling shorts in October. You want to increase sales; you have a hunch that if you that if you remove the sub categories you’ll improve sales. So in the original scenario the user goes from homepage>shorts>cargo shorts>blue 5 pocket cargo shorts>cart>checkout…
So you think it would be better to remove one step in the funnel and you do away with the subcategories and delete boardshorts, cargo shorts, walking shorts and compression shorts and lump all the shorts into one big shorts page with 80 different styles on it. In essence you go from a category that has 4 sub-categories that have 20 products each to 1 category with 80 products. If you make the change and your sales improve 5% over the next 2 weeks things went well, right? Maybe not; you can’t really tell by looking at the last step in the funnel. You may have had a decrease in the number of people that went from step 2 to step 3 in the sales funnel, but sales may have still gone up because you had an extraordinarily hot week in the middle of October. You may have negatively impacted sales even though sales went up. Know what your KPIs are for each step in your conversion funnel. Your analyst will be able to collect data specific to each step so you will be able to make assessments at each step. If they can help you improve every step by a couple of percentage points your final conversion rate will go up exponentially.
Additionally, an analyst will know how to segment and parse data to get a better understanding of how the KPIs are performing based off of a desired response and User Experience. I had worked on a project where we implemented heatmaps so we could see what people were clicking on and what they were looking at throughout the website. We implemented the system and started to collect the data. After a few weeks I pulled data to see how people were using a specific page and was happy to see that people were looking at the exact sections of the page that we had hoped. The page worked as designed! People we looking at the product data, spending time reading and interacting with the content and then they were clicking on the button to add to cart. About two weeks later I was sitting with an analyst who was talking about some major adjustments to that page because according to the analytics and heatmap, the page wasn’t serving its intended purpose. I disputed this because what I had seen showed that it was working; people look at content and then buy the product. She showed me that when you segment the users into people who purchase, people who bounce and people who continue on to search for something else, the page wasn’t actually working how we intended. A good analyst will be able to look past the initial data to make sure that they are telling the appropriate story.
Implementing an analytics strategy is tough, especially if your organization doesn’t really understand the intricacies of the different KPIs throughout a sight and don’t really understand what data they need and what reports will help them make improvements. A strong project manager should understand what kind of analytics strategy needs to be implemented and what type of skills an analyst should have to implement it. Building a project plan around implementing an analytics strategy can be tricky, but a strong PM will keep the main goals in mind and not let it get off track.