Competitive Intelligence: A Case Study

Gabriel writes in and asks:

nibbledish contains around 3,000 recipes and gets around 165k uniques per month, while sparkrecipes contains nearly 170,000 recipes but only gets around 16.3k uniques per month according to quantcast.  Do you have any explanation as to why this is the case?  I would expect sparkrecipes to garner much more traffic since it’s more established and it contains more recipes.

It’s an outstanding question.  Competitive intelligence is so crucial when analyzing any market, so I got Gabriel’s permission to write about the process I would go through to find out the answer.  I’m no expert, though, so feel free to add your own thoughts and/or correct mine.

My first step when researching a particular topic is to look at the sites in that niche and see how much traffic they’re getting.  I always use both Compete.com and Quantcast for that in order to get a more accurate picture, and Aaron’s SEO toolbar puts both of those within easy reach.

Trying to estimate traffic based on just a sample can be highly inaccurate, though, so don’t take information from either of these sites as absolute.  I’ve seen both of them at various times be off by almost an order of magnitude.  Normally I take the Compete.com number and double it, and that usually gets me close enough for what I need.

Sometimes you luck out, though, and find a site that lets Quantcast measure them directly.  That’s the case here with Nibbledish.  Looking at the Quantcast stats, we see that Nibbledish gets 389,000 visits/month from 331,800 unique visitors.  (I have Quantcast on a few of my sites, and they are very accurate when they measure directly.)

SparkRecipes.com doesn’t have direct measuring by Quantcast, though, so all we get is an estimate.  Time to see what Compete.com thinks.

Wow, Quantcast and Compete differ by a stunning amount — maybe more than I’ve ever seen.  Quantcast estimates 11,000 visits/month, while Compete.com estimates 352,879!  Who’s right?

Well, let’s assume that Spark and Nibble have the same types of visitors in general.  If that’s true, then Compete’s estimate of Nibble should be roughly as accurate as their estimate of Spark.  So since we know what Nibble’s stats are, we can use that to extrapolate Spark’s.

And thank you, Compete, for making the math easy for my Auburn friends.  For June they show virtually identical Unique Visitor stats for both sites.  That means that we can estimate that Nibble and Spark get about the same amount of Unique Visitors: somewhere around 325,000/month.

Here’s the really amazing stat, though (at least to me): Compete estimates that Spark gets 2.75X as many visits as Nibble does.  Same number of visitors every month, but Spark visitors come back almost three times as often.  Wow.

Related to that: Nibble gets roughly the same number of visits as visitors.  That tells me that people who visit Nibble usually don’t come back.  I also suspect, because of that 1-to-1 ratio, that most of their traffic comes from somewhere else rather than people going directly to the site.  People end up at Nibble from a Google search or a link somewhere, but no long-term connection is ever established so they never come back.  Spark is different, but possibly only because they have an entire network of sites driving traffic to the other sites in the network.

But back to the original question: why does Nibble get as many uniques as Spark, even though Nibble hasn’t been around nearly as long or have as many recipes?

It looks like the answer is SEO.  Take Compete’s referral stats with a giant grain of salt, but Nibble appears to be getting more traffic from Google than Spark is — both from Compete’s stats and from the 1-to-1 visits-to-visitors ratio.  Nibble’s homepage is a PR6, too, vs. a PR5 for Spark, so Google clearly sees more link love for Nibble.

An in-depth look into exactly how Nibble is winning the search engine battle would take another whole post entirely, but it’s obvious right now that their challenge is turning those one-time visitors into regular visitors.


Comments