Firefox 3, Download Day, and Why Browser Competition Pays Your Bills

Mozilla threw a release party for the whole internet last week. ‘Download Day,’ a coordinated push, put Firefox 3 on more than eight million machines in twenty-four hours and into the Guinness book for it. As marketing, it was superb. An open-source project generated Apple-grade launch energy on donated enthusiasm.

But the number that matters to the people who pay for websites is not the stunt figure. It is the quieter one underneath. On the client sites we measure, Firefox now carries fifteen to twenty percent of visits nearly everywhere. With academic and technical audiences, it is noticeably more. That share, not the download record, is what is changing your costs.

The browser itself, briefly

Credit where due before the economics: Firefox 3 is a genuinely better browser. Memory use, version 2’s chronic complaint, is much improved, and long-session sluggishness with it. Pages render noticeably faster.

The ‘awesome bar,’ their term, which has stuck to everyone’s mild embarrassment, turns the address field into a search of your own history. It reliably changes how people navigate within a week of adoption.

Under the hood, the Gecko engine’s standards support advances again. Better CSS coverage, better image handling, closer fidelity to the spec. For site owners the practical translation is pleasant. Standards-built sites needed essentially nothing fixed for this release, which is exactly what standards-based work is supposed to buy. Sites built to specification keep inheriting new browsers for free.

The share math, and what it retires

Now the strategic layer. Two years ago, ‘we only test in Internet Explorer’ was lazy but survivable. The neglected slice was single digits. At one visitor in six, it is negligent. And at the shares we see with university and research audiences, a quarter and climbing, it is a decision to ship a broken experience to people many institutions specifically court.

The support-matrix policy we have recommended since IE7 launch week handles this without drama. Measure your own audience’s shares. Tier your support to match. Revisit quarterly. What has changed is that the measuring now reliably promotes Firefox to the first tier. The organizations still treating it as an afterthought are usually the ones who have not looked at their own analytics since the policy was written.

Worth noting for completeness: this is not a two-browser story anymore. Safari rides Mac growth and the iPhone into visibility. Opera persists in its niches. The WebKit engine keeps spreading into pockets. The browser market is becoming a true ecosystem, which is exactly the condition where standards matter most. No one vendor’s quirks can define correct when there are five vendors.

Why competition is a subsidy

Here is the argument we keep making to budget owners, because it turns browser news from trivia into finance. Every website carries an invisible line item we have called the cross-browser tax. It is the hours spent making one codebase behave across engines with different opinions.

That tax is set by the least compliant browser with real share. For a decade, that was Internet Explorer, whose monopoly meant no market pressure to improve. What Firefox’s rise did, before any of its features mattered, was restore the pressure. IE7 exists because Firefox took share. IE8’s Acid2 commitment exists because the share kept moving.

Every point of share pushes every vendor toward the spec. Every step toward the spec shrinks the tax on every build and maintenance contract you will ever sign. That is how browser competition pays your bills. The eight million downloads were not just enthusiasts collecting a record. They were, in effect, votes for cheaper websites, cast by people who have never seen your budget and improved it anyway.

The four-engine board, updated

Where the field stands, for the matrix-keepers. Microsoft’s IE8 is in public beta, carrying the Acid2 commitment we covered in December. There is also a wrinkle the standards community is actively debating: how a more-compliant IE should handle the legacy content built for its bugs. That answer will shape enterprise upgrades for years.

Apple pushes Safari along the WebKit line: Macs, Windows nominally, and iPhones, where the share actually moves. Opera keeps its devoted slice and its habit of shipping ideas others adopt later.

And the developer-tools angle deserves its own sentence. Firefox’s extensions, Firebug above all, have quietly made it the browser web work happens in, whatever browser it ships to. That is a pull on the profession that share numbers undercount.

The scenario worth pricing for the planning horizon: a truly multi-engine world, no vendor above half, several above ten percent. That world is friendlier than the monopoly it replaces, but only for the prepared. It rewards spec-first building and punishes every remaining ‘works in our browser’ assumption.

What to do with this, practically

Three items for the checklist, sized to the news. First, the analytics look. Pull your browser shares this week, compare against your support tiers, and promote what the data promotes. Many of you will also discover Firefox 2 lingering. It does not auto-upgrade to 3, so expect a long tail.

Second, if your organization runs internal web applications, test them in Firefox 3 before your staff’s inevitable adoption does it for you. The intranet apps hard-wired to IE assumptions are the usual casualties, and June’s help-desk tickets are cheaper than September’s.

Third, and this is the one we would flag for anyone commissioning work: make standards compliance an explicit buying requirement, if it somehow still is not. In a five-engine world, ‘works in the browser we tested’ is not a deliverable. ‘Built to spec, verified across the current field’ is. Vendors who balk at the phrasing have told you something useful.

The longer view

Two years ago we opened this blog arguing that standards were the durable bet in an industry addicted to shortcuts. The evidence keeps arriving on schedule. IE7, then Acid2 commitments, then WebKit in pockets. Now the second-place browser breaks records partly on rendering quality.

The market is converging on the spec. Not from idealism, but from competition, which is sturdier. For everyone who builds on the web, the practical moral has not changed and keeps compounding. Write to the standard, test across the field, and let the browsers fight over who renders your correct code fastest. They are racing each other. You are collecting the winnings.