What Number is the Most Important? – Web Metrics (Part 1)

One reason the internet has become one of the most vital tools for marketers is that it allows them track to a level they were never capable of before. One of the problems I’ve had with this is that now we are almost inundated with data. More data at times than we could ever process, and in some cases more data than we would ever want to process. Really how do you deal with all this data? Ask two simple questions.

1. Does this data help me improve anything?
2. Does this data explain a cost or profit?

If the data answers either one of those questions for you then it’s worth looking at.  If not you are probably looking at something that is I’m sure very interesting, but ultimately a time waster.

Ernest WAS Right!

Hemingway that is… he once said “Never mistake motion for action.” Meaning just because you are collecting it doesn’t mean you’re taking action on it. When you start to put together your metrics dashboard think of the people who will be reading it.  What do you need them to do? You may want to have two, one for your web team, and one for your business team. Communicating the right metrics out to the right audience is just as important as collecting the right data.

The data your web team needs to do their job better is most likely going to be different than what the business will need to measure success. Usually though you will want your web team to see your business metrics, but not vice-versa. When deciding what to collect there will probably be two metrics that mean ABSOLUTELY NOTHING but will be the foundation of all your meaningful metrics, visits and page views. Visits should usually represent one users session in which they may have multiple page views. A page view or a site visit doesn’t really give you relevant information about your site unless you are looking to sell advertising space on it, then it only means something to the people looking to buy space (of course they aren’t looking at the right metric here either but hey you made your buck.).

What should I Be Looking For?

For the business side: It all comes down to how much you make or how many leads you generate or could generate. What bothers me the most is how reports are manipulated to confuse visits and page views. A site will report on page views (or impressions) if they are looking to sell advertising space, which is to their advantage as they are more likely to have more page views than visits (remember a visits consist of multiple page views), but then they will use the visitors number to calculate their conversion rate. This is actually how a lot of companies report their numbers for conversion rates and I’d argue that this is a rather poor way of reporting this IF you ever want to improve your conversion numbers. To me it’s simple every page is an opportunity to convert, not every customer is an opportunity to convert. You can actually convert a customer multiple times for certain actions (say adding more items to your shopping cart) or opting into other emails you have. If you start to think of every customer as a conversion opportunity you are leaving more than half your opportunities out there.
So here is what I recommend you look at for reporting the success of your business.

1. Relate everything to a dollar value.
2. Calculate your opportunities to convert based on the lowest common conversions element (page views when you can visits only when a single visit builds to one single action).
3. Web metrics collected by Google analytics or some other reporting software isn’t enough. Run surveys on your site.
4. Understand what if any offline influence you have.
5. Report honestly even if the news is bad.

So how do we do this?

Relate everything to a dollar value.

This one is simple, at the end of the day a business is about making money if you can show how your website contributes to the bottom line then you will see higher investment into it, but the responsibility is on you to establish this value and be consistent on your reporting. If you insist on showing the number of visits to the site or to a section of the site then please have a way of showing what this means in terms of sales. Does 1 million visits = 1 million dollars? What is your average order value, how can you grow that number? If you are going to show how many visits a section of the site has, what does this mean for revenue at that level? You should establish some basic formulas for calculating your revenue and revenue influence … we’ll get to that shortly.

Calculate your opportunities to convert based on the lowest common conversions element.

We covered this, but again if your opportunity is to convert a single user more than once during a session report conversion based on page views, if a conversion can only happen once during a session say clicking a banner that takes them away from your site, or completing a shopping cart check out then yes use visits but understand the difference between the two and report in a way that makes the numbers more accurate.

Run surveys.

This is the MOST valuable tool you have at your disposal and yet very few websites do this. You can run them in many different ways that are not imposing. You can have a simple survey with only one question and three or four answers under a navigation element, or if you capture an email address you can e-mail a survey out, or you can randomly sample your current online customers by popping up a survey. Anyway you do it the questions you ask should help to compliment and fill in holes your regular metrics can’y answer. If you aren’t sure how many people are buying your products offline, ask them in the survey online “do you intent to purchase the product you are looking for online, offline, I’m just browsing.” That question alone will give you a lot of information about your customers and your website. If you know your average online order value is $100 you can probably assume your offline is the same (not sure ask the customer how much they think they may spend to confirm this.) if 60% of your customers say they are going to buy offline then you know your website has made an impact to 60% of the population of your web users (remember they are the ones who answered this survey). You don’t have to worry about asking every customer, statistics and random sampling show that the responses of a few (sometimes only 100 people) can be enough to understand the larger group. I however like to make a few assumptions as well and guess that if 60% of people say they will buy offline only 1/3 of those people will actually do so (you can confirm this number further though if you are able to capture their email and send a follow up email one month later). This way you know at the very least you will be under reporting your numbers as opposed to over reporting.

Understand what if any offline influence you have.

I think we just covered this, to understand this you need to run surveys. What roll does your site play in a customer’s purchase path? A good site should play a roll in each phase of a customers buying cycle which are typically: Research, Selection, Purchase, Use, Evaluation, and Recommendation. However not all sites DO play a roll in each part of the cycle. You may only play a part in research, so where are they buying from? Time to ask in a survey. They take such little time to setup, but offer the most valuable return you can imagine.

Report honestly even if the news is bad.

This is the most important and most difficult item people have. Nobody wants to deliver bad news, but if you try and put positive spins on your numbers you aren’t giving people anything to action. People take action when things go wrong. So tell them what’s wrong. Look at your numbers and figure out what sections need to fixed and why. If you are reporting on something positive it should tie back to something unique that wasn’t done before and be used as a proof of concept this new way works. If you are reporting on growth but not because of anything you did on the web then understand what the marketing team did that may account for a spike. If you see an anomaly work with your other business partners to figure out why your numbers have changed outside of their normal patterns.

In part 2 we’ll look at metrics for your web team.