What’s more valuable to a company? A visitor to its website who spends 15 minutes scanning a wide variety of pages, or a visitor who comes and goes in three minutes? The obvious answer is the first one, because as any marketing executive can tell you, “stickiness” and time on site are drivers for the website experience. But what if the first person is taking so long because they can’t find what they are looking for and the second person came and left quickly because they readily found the white paper they wanted or even transacted? The lesson here is not that time on site isn’t the only metric you should be evaluating. In fact, using metrics to evaluate the performance of your site may not be as straightforward as it looks.

Take the recent news about Instagram over-taking Twitter in terms of volume last year. “Instagram Is Now Bigger Than Twitter” was the headline everywhere from CNBC to Re/Code to the New York Times. But how meaningful is that comparison? Twitter has some 284 million active monthly users, Instagram more than 300 million. Yet, as an article in Slate describes it, the two are different: “One is largely private, the other largely public. One focuses on photos, the other on ideas. They’re both very large, and they’re both growing.”

Another metric that is often bandied about is unique monthly visitors. This measures the number of people that come to a site and discounts repeat visitors. Again, that might sound like the ultimate metric for evaluating the attention that a site is getting. Still, it doesn’t measure what those unique visitors are doing on the site. If it is a content-driven website, like the Huffington Post or Buzzfeed, a more important measure may be “total time reading.”  There, the number of visitors who come and leave quickly isn’t very valuable to advertisers who provide the revenue for content-driven sites. Total time reading is far more important, and smart advertisers recognize the difference and factor that in accordingly.

A common measure reported on widely in the media when comparing different brands’ web traffic is the number of website visitors. This is frequently sourced to web measurement and analysis companies who make these types of evaluations. But even these can be highly misleading. First and foremost, according to a recent post in medium.com, the most widely quoted source of web traffic, Comscore Networks, only counts U.S. users. If a brand is global or operates overseas like a many government defense contractors, the metrics will not include that traffic in the totals. In addition, these reports are often based on sampling which can distort the actual numbers for smaller brands with a more limited number of visitors. It’s also not yet clear whether these services are including site traffic from mobile apps, which may be a very important measurement tool for many websites as more and more visitors use mobile devices to access information on the web.

So if the three most commonly-used metrics for measuring the success of a website—time on site, unique monthly visitors, and total traffic—all have their flaws, what is the best way to evaluate how a site is doing?

The answer is there is no best answer. All three of those key metrics are useful, but they need to be taken for what they are which is a set of imprecise and blunt tools.

A better way to look at the most effective mix of metrics is to find the best blend that will help evaluate “value.” Time on site is important, but only as an element in value. In reality, for media websites, advertisers don’t actually want a customer’s time, they want to make an impression that will lead to a transaction or buying decision. On the other hand, for an enterprise site offering IT solutions where the buying cycle is long and a visit to the website may be part of the research process, time is valuable as a measurement for a customer’s information gathering step in the cycle. Where they go on the site—to resources, for example—may say a great deal about where that customer is in the cycle and how to best to pursue him or her.

Where the visitor enters the site may be a key performance indicator for both organic search results or for a lead-generation driven campaign that takes the visitor directly to the intended content. Spending time on the blog page may be an indicator that the site’s content is fresh and engaging and is bringing target audiences back for more. Reading product and solutions pages may indicate a prospect that needs to be watched to make sure they are getting what they need to make a purchasing decision.

The right answer is that value has to be a combination of a number of factors, and using multiple metrics can help understand if the site is achieving its goal of providing that value. But no marketer should get too hung up on any single measurement.