How To Choose An Analytics Vendor: The Non‑technical Part


April 13, 2006

Yesterday, on the Web Analytics Forum, a visitor asked, “How do I choose a web analytics vendor?” JonB from England (who appropriately goes by the username of Mr_Awesome) answered with an excellent non-technical response. He skipped the usual issues of “Do you want server side or client side analytics?” or “Now what is it you want to measure?” and addressed evaluation of the company itself. In the email where he gave me permission to pick up his words wholesale, he added, “I learned all those lessons the hard way, of course.” Here is Jon’s list:

1. Look at their customer list. Are the businesses they serve in a similar market/vertical with a similar business model to you? If so that’s a good sign.

2. Ask to speak to their support/account management staff directly. Ignore the fancy sales guys as once you sign the contract, you won’t ever deal with the sales guy again (until it’s renewal time – at which time they will suddenly be your best friend again).

3. Never sign anything more than an initial 12 month contract.

4. Most importantly of all – check no less than 3 references that you choose from the vendor’s reference list. (Don’t let the vendor give you references… that’s just a pointless idea.) Better still, go out and find someone not listed on their ‘clients’ page and track down the person responsible for their Web analytics programme internally. Buy them lunch/coffee/chocolates and probe them on their opinions of the vendor.

5. Document your expectations from the tool and give this to them in writing before you sign.

My additions (outside of the techie stuff and the “what do you want to measure?” stuff) are:

A. What do you want to spend? I find that analytics pretty much fall into three categories:

  • Free/Almost free
  • Mid-range ($1000+)
  • Astoundingly expensive ($15-25K/year.)

Once you establish a price range for yourself, you’ll find that there are just a handful of reasonable choices.

B. Make sure that you are comfortable with the company’s commitment to their software — you don’t want to work with a new company who has few customers and who may easily go under in a year, or even a large company where the analytics area is not a strategic focus.

C. Narrow down your list and ask for a trial, free period with your top choice. You could even have a trial period with your two top choices and compare them.

Robbin Steif