⇥ How dense is your site?

October 9, 2008
3 comments

⇥ Permalink

How do you measure the success of your website?

There are a number of different answer to this question—mostly depending on who you ask. A programmer is likely to measure success by how well the site performs, or by how closely it performs to its original spec (assuming there was a spec in the first place of course). A business person is more likely to look at some sort of performance metric in the numbers: traffic, revenues, and so forth.

When I consult with companies who need help setting their web strategy, I usually find that many people have trouble finding an objective metric that they can use not only to determine the success of their website, but also to calculate its ongoing growth prospects and potential.

At MTA, we measure two things: how much it costs to run a website, and how much money it generates in relation to its traffic. Essentially, we look at our websites as a sales tool and, therefore, we evaluate them the way we would evaluate the performance of a salesperson—by how much it costs to sell.

The density of a website
The starting point for evaluating the performance of your site—past, present and future—is by measuring its density. We define density the way Internet ad agencies define the cost of their profit: by dividing some value by the number of pages a website serves. For example, dividing the total revenues generated by a website by the number of page impressions will give you its revenue density, while dividing the cost of running a website by the number of impressions will give you its cost density.

These figures are incredibly helpful in monitoring the performance of a website and managing its growth. Most of all, they are super-easy to calculate: all you need is continuous metrics on your traffic (which, thanks to tools like Google Analytics, are a trivial endeavour) and a mildly-competent accounting department capable of keeping track of expenses properly.

What should you consider?
On the revenue side, calculating the density of a website is usually pretty easy—if you sell ads, your ad revenue will be the base number. If you run an online store, your base number will instead be your sales. In some cases, things may get a bit more complicated: for example, if you use your site to drive leads to a sales team, you will need to ensure that you track revenues that are generated by the leads properly.

On the cost side, you will need to include both direct expenses—servers, bandwidth, etc.—as well as indirect expenses—salaries, advertising costs, and so forth. Some juggling might be needed for borderline situations, like for example, developers who work on the website and other applications, and so forth.

Understanding density
Density figures are more useful than the numbers that are used to calculate them because they provide a way of calculating the performance of a website in relation to its traffic; therefore, they give you a very good picture of how well your web business strategy is set up, and provide with useful clues as to where the areas of concern and improvement are.

Consider, for example, the case of a website that is primarily ad-driven. These sites are usually dependent on large amounts of traffic for their success and, therefore, you are likely to find a very low revenue density. Clearly, this means that the cost density needs to be appropriately small—otherwise, the company is not doing so well.

A company that has an online store, on the other hand, is more likely to have a very high revenue density, and a correspondingly higher cost density. For example, php|a has a revenue density of between $1,000 and $1,500 per thousand impressions. Here, too, the cost density needs to be smaller than the revenue density in order for the company to be profitable.

This, in itself, is not a particularly surprising notion: a company that spends more money than it takes in is not going to be in good shape. The importance of density is in helping you go behind these numbers and understanding what they mean:

First, it gives you a measure of the quality of your traffic. If your marketing campaigns are pulling in the wrong people, you will see your cost density increase, but your revenue density decrease. Thus, you can use densities to easily measure the effect of marketing and ad campaigns on your bottom line
Second, it provides a very good way to gauge your IT expenditures. Working densities in reverse from your budget, you will know how much traffic you need to achieve your goals, at what cost and with what conversion ratio between visitors and clients. At the same time, it will help you identify areas of cost that need trimming by breaking down the cost according to traffic—ie.: do you need more visitors, or do you need better programmers?

Density as a trending tool
If you have been collecting statistics about your site for a while, you will be able to trend your cost and revenue densities and see how they have changed over time. Correlating this with particular events will give you perspective over whether, for example, any given decision you have made that has affected your IT infrastructure has been a good one.

More importantly, you can use historical information to determine how growth will affect your costs and profits. For example, suppose that, right now, you need to spend $100,000 a year to serve 1,000,000 pages. This corresponds to a cost density of $100/kp. How does that figure scale? By answering that question—easily calculable once you divide your expenses between fixed and variable—you will be able to anticipate the cost of your growth.

If you are operating under a well-thought-out web strategy, you will find that the cost density of a website tends to decrease rapidly with an increase in traffic and then levels at an optimal point where your variable expenses, such as bandwidth, become the main driving factor behind your costs. Until you get to that optimal level, you should focus on limiting your fixed expenses, whereas beyond that point you will need to focus on your variable expenses.

If you are working with the wrong kind of software, you will find that growth is reflected in a cost density that trends upwards with an increase in traffic—and that’s a clear sign that you need to sit down and rething your web strategy before things go bad.

Either way, keeping a close eye on your densities enables you to track many important aspects of your web infrastructure—and, because they are so easy to calculate, they can do so almost in real time.

⇥ Some thoughts on the php|architect relaunch

October 6, 2008
9 comments

⇥ Permalink

It’s been a week since we “rebooted” the php|architect brand—time for some thoughts, lessons learned and selected statistics.

First of all, the reason why we did this. Many people have written me personally wondering what madness possessed me to authorize a price drop of over 50%, and moving to a colour-only, print & PDF format—some actually questioning my motives, as if the changes were really the result of an alliance with some sinister external corporate force that is silently taking over our company.

Reality is always much duller than fiction
From a certain point of view, during a period of questionable, government-sponsored bailouts that appear to reward incompetence and crime, I can see how people would welcome the news with suspicion. After all, when a box of pasta has doubled in price over the last six months, any price reduction looks strange at best.

Alas, there is no devil in the details. We decided that it was time for php|architect to reflect the market which it attempts to cater to—and the PHP market has changed dramatically since we introduced the magazine six years ago (that’s right, ladies and gents—six years ago).

When we launched php|a, I honestly had no idea that it would become the success it is today. We started because we, as developers, were unhappy with the content that was available, and needed a better source of knowledge—we never expected it to generate any significant revenue, let alone spawn the series of products that have come out of our labs over the years.

Meanwhile, the increased cost of producing the content of the magazine has been more than offset by a growing number of subscribers, and the price at which we were selling it was no longer reflective of the market we were reaching. $80 a year translate into a nice revenue stream, but that very revenue was making us drift from our mission of providing knowledge to the PHP world in a reasonable way.

Is the price right?
Thus, we decided that php|a needed a new image and a new price. Next, came the decision of what to do with our current subscribers. Technically speaking, we were under no obligation to do anything other than continue their subscriptions for the remaining number of issues at the old price. Or, we could extend their subscriptions in recognition of the fact that the price had changed so dramatically. We thought about it for about 1.539 seconds and decided to go with option B—any which way you look at it, it just makes sense. First of all, it’s the right thing to do. Second, from a business perspective I doubt that customers would have been very happy with us if we had done nothing. Third, from a financial perspective the cost of handing out 30,000 free issues is amortized over time, and will be offset by the advantages of having more long-term subscribers.

The decision of going to an all-print format was much more agonizing. In the end, we considered that a larger number of print subscribers would afford us the opportunity of amortizing certain costs among a larger pool of customers, while providing us access to products, like colour printing, that were simply not available before.

To print, or not to print
A number of people have written us with concerns (“concerns” being, in some case, a mild euphemism for death threats) about the environmental impact of this decision, which is something that we did, of course, take into consideration. Although we did mention, in our press release announcing the new php|architect, that the changes would essentially be carbon-neutral, that statement just didn’t seem to compute all that well, so I thought I’d pass along a slightly expanded explanation here.

The print industry works in very strange ways. Despite the fact that immense strides have been made in the technology behind printing, the actual production of printed materials relies on processes that haven’t really changed in many years—aided in part by the fact that the capital costs involved in running a printing operation are enormous.

Many people I talk to seem to think that magazines are produced the same way we go about printing pages in our day-to-day lives—print the page one by one, collate them, and ship them out to the customers. They imagine bigger printers, but essentially expect the process to be the same.

Sadly, that’s not the way things work. Large-scale printers have a number of quirks that are not immediately evident to the uninitiated individual (as we ourselves found out the hard way when we started printing php|a). For example, lighter paper weights are only available on large web presses, which, in turn, are only economical over a certain number of copies due to the significant start-up costs caused by capital amortization, mounting the plates, inking them, and testing ink distribution. Similarly, environmentally-conscious products like acid-free paper and soy-based inks are only financially feasible at larger quantities just because of the way they are priced.

Thus, printing, say, 10,000 copies of a magazine in colour can cost barely 20% more than printing 5,000, but result in a final product that is lighter—and therefore has a much lower transportation carbon footprint—and is made with materials that are much better from an environmental perspective.

So, what did we learn?
By and large, we found that our customers have liked the changes we introduced, resulting in a jump of almost 20% in subscriptions in just over a week. It is still too early to tell whether the trend will keep up, but so far I am very happy with the results.

One important lesson that I learned from this process is the significance of change. Despite the occasional complaint, the general message that I got back from our readership has been a collective sigh of relief, almost as if they were trying to tell us “about time!”. Well, there are more (big) changes coming, so I look forward to more and more occasional audible breaths from our audience in the near future.