My background is in ecommerce and Internet marketing. Over the years, I’ve used lots of online marketing channels (SEO, PPC, shopping engines, affiliate programs, email, display/re-targeting, social, etc.) to drive sales. I call myself a metrics-oriented or data-driven marketer because I base a lot of my decisions on the data I collect. As many people have said over the years, if you don’t track results, you’ll never know if you’re succeeding or failing.
Leveraging powerful analytics solutions like Google Analytics, Webtrends, Coremetrics, and Omniture over the last 15 years, I’ve attempted to track everything possible and made decisions based on a number of metrics.
In the ecommerce and Internet marketing world, there are general metrics which are interesting like page views, unique users, average order value (AOV) and marketing channel-specific metrics like open rate, click-through rate, bounce rate, unsubscribe rate, etc. Then there are meaty metrics like conversion rate, which combined with other metrics, lead to the bottom line: cost per acquisition (CPA), lifetime value (LTV), return on investment (ROI), return on ad spend (ROAS), etc. More recently, I ran a SaaS business called SingleFeed and started looking at retention waterfalls and cohort analysis as well as everything that our CRM and marketing automation systems had to offer.
In other words, there are lots of metrics running around in my mind.
And then I started working with Kontagent, and I was introduced to how gamers think about their world. It took me a while to grok the new world order of Farmville, The Godfather, and The Sims Social, but once I did, I realized that there’s a huge opportunity for ecommerce companies to take advantage of the metrics which drive social gaming upstarts.
As Dan mentioned in a recent post, it’s not about page views anymore. While I disagree with some of his points–CTR is an extremely important internet marketing metric!–the gamers are on the cutting edge of analytics. In particular, their use of cohorts, engagement & retention, and event timelines are important drivers of success, which will start to bleed into general Web metrics. Etailers need to start leveraging this type of data today or risk being supplanted by upstart social commerce companies.
Here’s a quick peek at some gaming and mobile metrics which etailers need to pay attention to:
Most etailers I worked with before a couple years ago didn’t think about cohort analysis. Many still give me a blank stare when I bring it up because it’s just not yet part of the standard lexicon for ecommerce. You can look up cohort in Wikipedia to get a more formal definition, but in my own words, a cohort, and more importantly, a cohort study or cohort analysis looks at a specific set of users over some time period. Looking at cohorts as opposed to everyone can help a company make better informed decisions.
For example, at SingleFeed, we looked at customers we acquired in different months and what happened to those customers over time. What was the retention rate of the users we acquired in January vs. February vs. March? In this way, we could determine that one cohort–let’s say the February cohort–had a higher churn rate compared to the January cohort. This would then push us to ask ‘why?’ and dig into those acquisitions. Was it because of a marketing tactic? What it because we brought in a different customer demographic (Industry? Size of company? Revenue?)? Was there an internal SingleFeed change (Pricing? Feature? Process for on-boarding?)?
With Groupon’s IPO, a lot of people are discussing their cohorts. This blog post at Techcrunch questions the profitability of Groupon’s cohorts over time. Looking at Groupon’s acquisitions in aggregate would just show that the revenue from their membership is growing and growing, but looking at cohort analysis might show that newer members aren’t spending as much (and therefore less profitable) as quickly as older acquisitions.
Etailers need to do similar analysis. Here’s a snapshot of customer acquisition broken down into weekly cohorts:
While the chart talks about installs as opposed to members or customers, the point is that it’s no longer just about spending $X to generate $Y in sales. Retailers should think about the lifetime value of the different cohorts and adjust their acquisition marketing accordingly.
Engagement & Retention
Ecommerce companies don’t typically think about engagement and retention, but there’s a new breed of retailers called flash sale sites – think Gilt Groupe, Zulily, and Lot18 – whose business live and die on engagement and retention metrics. These companies spend a good amount of money to acquire a user and don’t always make their money back on the customer’s first purchase. Even if they did, these companies wouldn’t be the powerhouses that they are unless they could get their customers to come back, re-engage with the site, and make repeat purchases. In this way, the flash sales sites are thinking about the lifetime value of their members.
Regular ecommerce sites have to start thinking similarly. No matter what you’re selling – shoes, tires, tables, or beds – your competition on the web is just a click away. What you need to do as a retailer is engage your customers and encourage them to return to your site for more information about the products they bought, ancillary products they might be interested in, and introduce them to a community of users who can help them in some way, etc. In this way, your current customers will be more likely to buy from you in the future. This is part of the reason smart retailers set up blogs, create a following on Twitter, and buy Likes on Facebook. It’s always cheaper to re-market to your current customers than to acquire, educate, and win the hearts of new customers.
The current “standard” of Web analytics have some information about engagement and retention, but it’s not good enough because it’s based off of cookies. At Kontagent, we’ve started to work with ecommerce sites that have persistently logged in states through Facebook connect, for example. And therefore, we can provide true engagement and retention data.
Here are a couple relevant snapshots from the Kontagent dashboard:
If you run an ecommerce site, do you know what percentage of your customers are coming back on a daily basis? Do you track changes in engagement and retention in order to pinpoint what drives these metrics? Are you able to look at daily, weekly, or monthly cohorts? If not, it’s time to start.
Gaming companies think about every action that takes place in the game. Every time you click on that pig to feed it, click on the castle to upgrade it, or click on that quest to get to the next level, the gaming company logs and analyzes all that information in order to help them figure out how to improve engagement, retention, and monetization.
But general Web companies don’t think like this. For an ecommerce site, the standard is to get lots of traffic, improve conversion rate by setting up better landing pages, and collect orders. That’s way over-simplifying things, but ecommerce sites do not typically think about which events or actions that take place on their site are really important or which actions lead to other actions.
Without an event timeline, you’re just making a lot of assumptions. With an event timeline, you’re making decisions based on data and can therefore significantly improve your business.
For example, a sporting goods retailer might see that users who purchase from the rock climbing category are more likely to purchase from the outerwear category next. Because of this, the retailer can create custom marketing messaging to properly up-sell the right category to the right users at the right time. This type of business intelligence is standard in social gaming, but sorely lacking in the ecommerce industry.
And you can take this one step further to the product level. With an event timeline, you can see that users who purchase running shoes are more likely to purchase running shorts within the next ten sessions compared to the average user. Why not push those running shorts to the user with a personalized site experience or a one time coupon?
Here’s an example of what an event timeline might look like for a retailer with six different product categories, called purchase_item, purchase_item2, purchase_item3, etc. The X-axis is the session number:
This is one of the most powerful tools for gaming companies and should immediately become an essential way to think about your ecommerce site. You need to understand not just the basic funnel which gets your customers to convert, but what they’re doing over their lifetime in order for you to maximize LTV.
There are many other gaming metrics which retailers are going to need to understand to drive their business in this social and mobile world, but cohorts, retention & engagement, and event timelines are going to be at the core. If you work for a merchant that is still stuck in the world of old school metrics, I’d highly recommend doing all you can to understand the metrics behind social games.
Zynga did over $300M in revenue in Q3, which would put the company on track to be a $1.2BB revenue business. That rivals the online sales for Gap, Williams-Sonoma, and HSN, #24, #25, and #26 respectively on the Internet Retailer 500 list. While the average retailer isn’t going to start selling virtual goods and create the same social experience as Zynga, it couldn’t hurt to understand the underlying metrics which have driven Zynga’s meteoric rise over the last couple years. Those same metrics might just be what drives growth for your ecommerce business over the next couple years.
About the author: Brian Smith is an ecommerce consultant for Kontagent. Most recently, he was an analyst for ComparisonEngines.com and CEO of Singlefeed, a tool that simplifies the complexities of comparison shopping engines so that online retailers could easily and cost-effectively sell more. You can contact Brian at: @brianasmith.