Two Steps Forward for Online Brand Measurement

Last week two announcements brought renewed industry attention to the long-suffering field of online brand measurement. Eyeblaster and Dynamic Logic partnered to measure the impact of rich media ads, and InsightExpress introduced a change to its survey recruitment methodology. Each of these announcements promises to bring brand measurement ever closer to becoming a staple of online ad reporting.

When advertisers measure brand impact, they can actually test any or all of the following factors:

• Unaided brand awareness measures how many users remember an advertiser’s name without being prompted (e.g., Name five software companies).
• Aided brand awareness measures how many users recognize an advertiser’s name when it is presented to them (e.g., Which of the following software companies have you heard of? Microsoft, Adobe, Nate’s Software).
• Ad recall measures how many users remember seeing an ad for a specific advertiser (e.g., Have you seen an online ad for Nate’s Software?).
• Message association measures how many users remember the key messages that are included in an ad campaign (e.g., Which of the following best describe Nate’s Software? Easy to use, inexpensive, stable).
• Brand favorability measures how many consumers view an advertiser’s brand favorably (e.g., Is Nate’s Software a cutting-edge brand?).
• Intent to purchase measures how many consumers are likely to purchase an advertiser’s product (e.g., Do you plan to purchase products from Nate’s Software?).

One reason I doubt ad display time is an effective proxy for brand impact is it can’t measure any of these factors individually. Instead, ad display time tries to paint a picture of brand impact with a single, broad brush. To paint an accurate picture of brand impact, you must survey consumers.

Control/Exposed Method Vs. Pre/Post Method

Online brand impact is most commonly measured using control/exposed surveys. This methodology compares an exposed group of users who have seen an ad with a control group of users who haven’t. The exposed group is surveyed immediately after the ad has appeared on their screens — they are typically notified by pop-up windows that they could win a small cash reward for participating in the survey. The control group is recruited on the same sites in the same way during the same period of time.

Less common is pre/post measurement. In this methodology, a sample of users on a site is surveyed before and after an advertiser’s campaign runs. If the campaign is successful, the brand impact numbers will be higher in the post-campaign survey than in the pre-campaign survey.

Each methodology has advantages and disadvantages. Control/exposed surveys can provide statistically significant results with only several hundred respondents. Advertisers can also gauge the impact of their campaign by frequency, helping them determine how many times their ads should be shown to each user for maximum brand impact.

A control/exposed methodology doesn’t measure the effect of an entire campaign, though it typically measures the impact of a particular ad within a campaign. In addition, running a control/exposed survey on each creative would be cost-prohibitive for advertisers and would overwhelm sites and users.

Pre/post surveys more closely resemble how offline advertisers measure the campaign effectiveness. This methodology tells advertisers how the entire campaign impacted an entire group, not just how a few creatives impacted a few exposed users. Pre/post measurement typically requires a larger number of respondents, however, and therefore often costs more than control/exposed measurement.

Eyeblaster and InsightExpress Drive Ahead

The new partnership between Eyeblaster and Dynamic Logic makes control/exposed brand surveys cheaper and more accessible. Advertisers running an Eyeblaster ad will now be able to add basic Dynamic Logic brand impact studies to their campaigns.

Advertisers can set up the preformatted Eyeblaster Brand Test within the Eyeblaster trafficking tool. No need to call in experts to run the survey. The $5,000 per-test price shaves several thousand dollars off Dynamic Logic’s next-cheapest offering, making brand testing accessible to a greater number of advertisers.

When the Brand Test formally launches in January 2004, it will be the only private-label offering from Dynamic Logic. It’s not a perfect solution. For instance, the price tag is still a bit too high to encourage regular usage. But it will certainly offer more accurate brand measurement than the current crop of ad display metrics. If the program proves successful, it will only be a matter of time before other rich media firms and online ad servers offer similar products to their clients.

Meanwhile, one of Dynamic Logic’s competitors decided to use floating ads in a different way. InsightExpress announced it will start recruiting survey participants with floating ads rather than pop-up windows.

The firm believes this change will create a better user experience and obtain a more representative sample for its surveys. In the face of the constantly growing crop of pop-up blocking software, it’ll also ensure InsightExpress’ ability to recruit panels as pop-ups become less viable.

Both announcements will push online brand measurement ahead and get more advertisers to measure their campaigns in this way. The objective, of course, is to make online brand impact as easy to measure as click-throughs. Hopefully these developments have brought us one step closer to that goal.