The inspiration for this post comes from the High Scalability blog, which has an article on how your analytics product is lying to you . A few weeks ago, I was forwarded the link to the Walmart case study on their performance optimization journey. Highly recommend reading through the whole thing and really thinking through the implications. If you have been reading my blog for a few months, you might recall this post, in which I suggested that all your web page speed benchmarks might be wrong. My inspiration then was the 37signals blog, which is an incredible read. Also strongly recommended. Before I go further on the topic of this post, I owe a LOT of inspiration, education and understanding to my new habit of reading Hacker News on a regular basis.

Now that I’ve paid the “link debt” I owe to the greater web, I’m going to focus on what I’ve learned and the implications of how we think about building, structuring and managing the “performance process.” In thinking about website performance over the past few months, and reading about how other very smart, incredibly talented people are also thinking about the same topic, I’ve begun to embrace some bigger picture concepts as it relates to performance. Granted, this framework might not be new to marketing folks, tech folks or design experts but it’s a new way for me personally to think about the initial investigation work, to the progress and finally to the outcome of “World Class Performance” (™).

Will the future we live in be slower or faster than the past?

Part of the process that goes into investment decisions is the balance between short and long. This is a standard conundrum, and why many large firms get disrupted by startups. Change the game, and you have new rules. Change fast enough, often enough and pretty soon, there are rocket ships launching, flying and falling at a regular interval. The rules left behind by these successive launches become the new standard for those rare companies that launch, fly and last. I believe, as anyone who has studied the trends of the web should believe, the future is faster than the past; faster internet speeds, rising expectations and greater consumption online. That was the thesis of my previous post on this topic, which was that expectations will rise so not only do you need consider performance optimization today, but also plan for the future. One nice thing about performance optimization, from a prioritization and marketing perspective, is it has immediate gains in conversion rates, as well as longer term impact on innovation, reliability, trust and value.

Let’s get down to the principles and the stages, based on my observations of the space and where we’re headed. It’s an amazing place, full of fast loading, intuitively designed and inspiring experiences which prompt the natural desire to share the joy with others. Are you with me so far? Good.


These basic principles have helped me define the stages of performance optimization, the business implications and the process by which you should engage the team, design your strategy and assess your results.

1. Speed requirements increase over time as expectations rise – “fast” is a relative metric, not an absolute one. Thus as more companies get “fast” those that want to continue to be perceived as “fast” have to push the envelope.

2. Prospect modality, or the way in which the addressable market behaves, is dictated by their a priori expectation. It’s safe to assume your potential customers are doing three things online: checking email, searching and social networking. Each of those activities has essentially 90% penetration, or will by the end of 2012. So if your site differs greatly from one out of these three modalities in email, eg, using Outlook, Yahoo! Mail, Hotmail, Gmail, AOL Mail, or searching using Google / Bing, or connecting with friends & family on Facebook / Twitter, odds are 100% solid you’ll fail. Disruption occurs in two standard deviations from the norm, which I’d suggest is one part commercial innovation and one part true disruption, reframing the value proposition or changing the source of value in a given context. One specific example of this: links are blue. Period. Look at Facebook, Amazon, Apple and Google. Links are *blue*. If you design elsewise, you’re creating dissonance, adding cognative load and essentially making it hard for your potential customers to do business. Personally, I abhor the idea of a company imposing “creativity” on me because their designer couldn’t be bothered to make stuff as clear as possible. Links are blue is only one specific example of a list of design principles.

3. Sharing is a natural extension of a great product experience. I covered this reasonably well in a prior post. It boils down to setting expectations and delivering well above the water line. Many companies have started to formalize “Net Promoter Score” as a way of measuring their success or failure. Another favorite metric is “Sentiment Analysis” to figure out if the chatter is good, bad or neutral which accompanies your brand online. Sharing gets baked into so many different things in so many ways, I want to be clear about the sharing in this list, which is that sharing is the act of telling somebody else, in person, about the wonderful story you had, staring a product which fulfills some need. It’s not clicking a like button, retweeting or other vanity metric. It’s inspiring the good old fashioned phone call, daughter to mother, or son to father, to tell them about the time you first encountered “product or service Y” and how it blew your mind / fascinated you / was better than sliced bread. You get the idea. The product that wins is always the one most share worthy in “reality” not just online.


Still with me? Awesome, because I’m about to give you the keys to the universe, at least in terms of performance optimization. Here are the various stages, characteristics and some business implications of each part of the transformation from beginning to end:

Stage 1: Complete ignorance. No dashboard, no understanding and no reporting. This stage represents the majority of companies today, including Bank of America, (my bank, which I find especially deplorable, because they take my money and willfully waste my time), (I would expect the most commonly used operating system in the world to better understand the value of their customer’s time), and (Berkshire Hathaway, are you listening? All that money spent on television ads is NOT converting as high as it should because you guys are missing the foundation with which to build your digital empire. The reality is the consumer experience on each of these websites is much, much slower than it should be. So I give them all an “F” grade.

Stage 2: Proud of their Yslow and strutting like peacocks. These are the firms that have performed “plan pass one” and did some work. I would put FunAdvice, the company I co-founded, in this category. In 2007, we did the first performance optimization project and dropped our bounce rates by 10%. That’s a huge deal; user metrics rose across the board. However despite our monitoring since, we never measured the right things. The fallacy of measuring the wrong numbers or not having a holistic view and continuing to ask the hard questions results in stagnation, decay and ultimately disruption of your core business by those who “get it.”

Stage 3: Leveraging Yslow, httparchive and still missing the bigger picture. If you haven’t seen this site yet, do yourself a favor and read it. Even if you don’t read much, please, please read through the stats on this page. The web, overall, is getting slower. The quote, “we have met the enemy, and he is us,” is especially appropriate in this context. Suppose you have your dashboard, you have the third party industry trend data in hand. That’s the point at which you know two things: your stats, your market stats, and exactly how you can drive sustainable competitive advantage. This principle was the first in my list, above, for this very reason. Understanding trends is only as valuable as your ability to *act* on those trends, influence the team you work with and drive for those results you crave. Without that stitching together, companies at this stage get some benefit but not the long term halo of the second principle, that of prospect modality. Companies in this category have a more holistic dashboard than the companies in stage 2, but still miss the bigger, more important picture around principle two, which is that your digital presence will always be compared to the sites your audience interacts with regularly, prior to coming to yours. I don’t use Bank of America daily; I use Google, Facebook and Amazon much more often. These companies “get it” and clearly have embraced a holistic scorecard, shown their own case studies to the world and pushed all of us to embrace the idea of a faster internet. Godaddy (yslow score of “B” why not “A”), Digg (remember them? Grade “C”), and (was Mixx, now part of Bill Gross’s latest global domination strategy – Grade C). I’m positive the smart, motivated and savvy marketers and programmers behind all these web based companies know performance, know KPI’s and know their relative success vs competitors. However, in absolute terms, none of them are scoring a goal, which is the main reason these firms and in this category. In other words, for these companies, they are happy they are good on Yslow and beating their category, but the main, missing ingredient is that they’re not benchmarking against the overwhelming majority of users time. Facebook, Amazon, Google and or Apple. If you aren’t as fast as them, you’re slow, even if your site is the fastest by a mile in your category.

Stage 4: Benchmarking against the best in class, destroying your category and feeling good. The challenge with this stage is your company is now in rarified air. The performance metrics are in place, all fully green and compared to any other firm that sells the same variety of widgets, you are head and shoulders above the competition. However, the companies at this stage are not the Amazon, Google or Facebook’s of the world. TripAdvisor would be in this category, they’re clearly head and shoulders above the competition and their overall stats are very impressive. When you benchmark this kind of result to, which is a similarly complicated home page, the trend is clear: TripAdvisor is world class in it’s own right. However, have they built the fanatical dedication to performance into their DNA, have they broadcast far and wide their commitment to these KPI’s and made significant progress? In some ways, yes, as shows on their overall page size, javascripts and CSS they made a significant investment about six months ago. Interestingly, Quantcast seems to show a massive step function change in their audience and usage, tying into their performance stats: The trouble is, despite their solid performance, TripAdvisor is not one of the best performing websites in the world. At their scale, Tying Amazon or edging them slightly is not enough. They need to set a new standard, to force the industry to catch up to their level, rather than settling for equality among the industry leaders of today.

Stage 5: Setting the standard and continuously pushing the envelope. This is going to sound controversial, but I would put 37signals in this category for one specific reason. Ruby on Rails backend performance has become possible in the 50-75 millisecond range, which compared to a decade ago, is an incredible milestone. As DHH is one of the principles in 37signals and their Basecamp redesign is a case study in amazing, they deserve the credit from the front and back end performance benchmarks in being truly forward thinking and innovative. Google also deserves a ton of credit, as you read in my earlier blog post referenced above, for their bold 2009 study. Without the courage to put a public stake in the ground, it’s possible none of us would realize how problematic it is to systematically waste people’s time. Those companies that have the courage, the long term view that performance matters and signal to the rest of the digital world that wasting time, reducing productivity and creating discomfort are unacceptable aspects of an online experience are those which have reached this stage.

Selfishly, this post is entirely geared towards motivating every large, commonly accessed websites to become a stage five on the performance progression scale. Personally, I think it comes down to principles and strategy. It’s really hard to say, “We’d like to waste people’s time, decrease our GDP and create more churn,” in any meeting, large or small at a company with sufficient scale and influence. However, the sad reality today is that Geico, despite their advertising slogan of “we can save you 15% or more on car insurance,” has a fundamental breakdown between that marketing campaign and the commercial reality. Instead, they should say, “15 minutes can save you 15% or more on car insurance; it would be 10 minutes but we don’t value your time enough to care.” Yes, that’s a bit finger pointy but long term, should we, as a society, value highly companies which consciously or unconsciously waste our time, energy and inhibit productivity improvements? My vote would be a resounding “no,” that we should embrace these principles and both understand where on the performance ladder we are, and how to reach a loftier place.

Share →