|EPV - What Is It?|
... and why should you care.
| 3:42 pm on Apr 25, 2001 (gmt 0)|
EPV = Earnings Per Visitor
From a financial standpoint, I have seen many gauge a site's value by number of pageviews (impressions) and by total revenue. IMO, the best way to rate a site financially, is by number of unique visitors and your EPV.
My reasoning - you can have a sticky site that generates many pageviews per visitor but no revenue, or a site that has very few pageviews, but makes money. So, a site with 1K visitors per day sounds great, but if the site makes less than $10 per day, the EPV could possibly be improved. However, a site with 100 uniques per day, making $10 per day, is much more valuable due to a higher EPV.
Unique visitor = 1 pair of eyeballs = 1 shot at revenue, regardless of how many pageviews.
Take one month's site revenue, less operating costs. Divide that number by the total number of unique visitors, and you have your sites EPV. So, if site xyz.com made $5000.00 for March with 30,000 unique visitors, 5000/30000 = .166 - round that up to .17 for a $.17 EPV.
With this figure, you can safely estimate ROI on marketing, buying traffic or SEO services. You can also track your sites EPV from month to month, and look for patterns and warning signs. Regardless of your traffic fluctuations, hopefully your EPV will remain somewhat constant. (Seasonal sites excluded.)
Knowing your EPV before and after any changes can help you gauge your work. If you change your site layout, navigation, colors, etc. you should watch your EPV to determine if your changes helped or hurt. If you decide to launch a PPC SE campaign - knowing if the new traffic hurt or helped your EPV is golden info.
What is your favorite way to measure a site's financial performance?
| 3:52 pm on Apr 25, 2001 (gmt 0)|
>What is your favorite way to measure a site's financial performance?
Same as above.
Costs X 2 = Minimum Profit Goal
| 5:15 pm on Apr 25, 2001 (gmt 0)|
Very good logic, Drastic. Smart thinking.
| 5:27 pm on Apr 25, 2001 (gmt 0)|
Thanks for the tip Drastic, that's a great way to evaluate a site and to track its performance.
"Take one month's site revenue, less operating costs. Divide that number by the total number of unique visitors, and you have your sites EPV."
This, however, places great emphasis on having a good tracking program. One issue that I'd have with this is the quality of unique visitors. Our tracking software, Urchin, shows several discrepancies that we're still unable to account for. For ex., for most of our sites, at least 50% of our traffic is comprised by "no referrals"...and those are counted as uniques...but who are they? Definitely not people typing in the domain or coming from bookmarks (they're promo domains)...spiders & syphons - that many??? Another issue would be the AOL proxy..my understanding of that is that it can skew the number of uniques. Also, are these "high quality" visitors or people who saw an untargeted ad?
I guess the best way to take this into account is by tracking "EPV" through time. So once you determine the EPV and track it for a certain amount of time, you will hopefully determine a pattern...and a way to gauge for success/failure would be to observe trends and drastic changes to the EPV.
Since we all use different tracking programs, we need to know how it operates, observe change through time, and measure accordingly.
| 6:16 pm on Apr 25, 2001 (gmt 0)|
>places great emphasis on having a good tracking program.
Very true, know your log-analysis program if you want this to be accurate. Mainly, know how it determines what a unique visitor is.
Once you figure out what the log program considers a unique, then you need to filter out spiders and the like. If the site has at least a few hundred uniques daily, you can get an average % to knock off of your log tool's unique count when run without filters.
As far as referrals, it doesn't really matter where they came from. Browsers leak referrals, block referrals, etc. As long as they are making html and image calls, you should consider them a unique. Also, it doesn't really matter which of them are more targetted - you are looking for an average of all visitors.
AOL proxies are a different story. One thing I have done to try to get a better picture of my uniques is to use one of the free image/JS based counters for a few days. Some of these produce seemingly accurate numbers for AOL users, and can give you an idea of what your log tool may be missing/misinterpreting.
>tracking "EPV" through time
I think the best use of EPV is for snapshots of time - like a month or quarter. You want to take that data to gauge against other months/quarters to identify movement in EPV. This allows you to spot strengths and weaknesses that you can work with.