9 сентября 2015 г.
For a company to be as successful and competitive as possible, business owners need good data. Good data allows the canny business strategist to understand their customers on an unprecedented level - if they know what metrics to follow and how to interpret that data.
While digital marketing and web analytics offer deep, useful, actionable insights, many common metrics are like a net with holes that are too big. Potentially useful information just slips right through the cracks. Data from apps, software, and the cloud is often overlooked and doesn't make its way into future business plans.
Cedexis is an analytics company that specializes in helping ecommerce websites increase their conversion rates. We talked to Cedexis' Pete Mastin about some metrics that often get overlooked but are particularly useful for overcoming KPI blind spots.
Can you talk about how traditional web metrics are not giving the whole picture, leading companies to miss out on key business opportunities?
A performance indicator is just that: an indication of performance. What you are calling performance can really be anything. We could measure page load time and decide that's the most important metric; or we could dig in deeper and decide that for a particular business, 'conversions' are the most important metric. Those are business decisions, not technical or operational. They need to be determined by the business owners.
Often, web companies get lost in the "bottom up" metrics that are commonly discussed without understanding or making the causal relation to the really important business indicators - whether that is revenue, reduced churn or something more complex. To a large degree, this is why Cedexis introduced the "Impact" product.
Another KPI blind spot is SaaS (software as a solution). Do you have any advice on how to measure SaaS' success - in either their own or their competition's case?
SaaS is becoming the de facto way to deliver software. Other than mobile, almost all software being developed today is SaaS. There are exceptions to every rule, of course; but if you and me had a sharp idea about how we could develop some new software product, you would have to really work to convince me to write it as a shrinkwrapped CD.
Because SaaS runs on a best effort network that has known micro-outages every day, availability and performance have to be something built into the core of the product. Furthermore, there have been many studies out there that show that if you have a PLT of more than 3 seconds, you start to get serious site abandonment. If it's video, then your video start time had better be under 4 seconds. Performance is what keeps a lot of SaaS companies from succeeding - even when they have a really good idea.
To get the best possible data, business owners must decide on the best metrics. What differentiates a very useful analytic from an overly broad, and therefore useless, one? Can you give a real world example of this?
I think page load time is a really good example of that. I am kind of a baseball nut, so I once wrote a blog about how page load time is kind of like batting average (BA). Batting average as a metric has been shown to have huge flaws, and that is why sabermetrics came into existence (so now you have a whole new set of metrics around hitters): to overcome these older metrics. PLT is like that. There are newer metrics related to user experience that should be looked at carefully to understand how they might affect your specific business. For example, in the blog I talk about how one of our customers found more value in successful page loads (SPL). You can read about it here:
Many metrics measure only in binary, either/or thinking - either a customer is on your page, or they're not. Can you talk about the difficulties in measuring more vague concepts, like customer satisfaction or probability of returning
One really interesting case study we have is from a consortium that wanted to have a "Green Cloud." They had about 10 private clouds and they wanted to federate them and use the "Greenest" Cloud when performance was equal. It turns out there was already an organization that rated data centers on their greenness, and we were about to use those KPIs to select the correct cloud to route traffic to.
You can read about UnivCloud here:
How can these insights into traditional KPI blind spots, especially when performed in real time, give a leg up over the competition?
I think people need to think very broadly about KPIs. If you just use the ones that everyone else uses, you have to compete on the performance of those KPIs. But if you are thoughtful about your business and what will make it successful, you can create an entirely new set of metrics and measure yourself by those. Of course, in the end, a business needs to make money; but exploring the relationship of your business and the parts that make it up is what any thoughtful business leader should be doing.
For even more insights into your network and software's performance, request a free demo of TruView today!