Response data is invaluable for targeting those with specific interests, life-stage events, and other demographics as there is often no other way to obtain this data. Response data is primarily gathered from surveys and product registration cards and is often enhanced with traditional compiled data yielding highly selectable, targeted lists.
1.Self-reported information tends to be exaggerated and is the source of a great deal of “junk mail.” Human nature often takes over and those completing surveys and product registration cards tend to overstate their income, understate their age, and exaggerate their interests and occupations.
In many cases, the responders have incentive to respond to receive coupons, web-site access, or some other free or discounted products or gifts, but have little reason to respond accurately.
2. Every available check and balance should be used to insure the responders match the targeted demographics.
As an example, while someone may check a survey box indicating an interest in international travel or food and wine, making them appear to be good prospects for high-end products, many will not have enough income to purchase high end goods.
While you can eliminate some unqualified prospects by selecting higher income ranges, the income is often also self-reported on the same survey and it may be (is often) overstated. Here a second select like net worth or income producing assets can eliminate many of the unqualified prospects. These data elements are not self-reported and are instead modeled by experts such as Claritas and CACI using housing, census, and other public record data.
3. Beyond self-reported information about interests and lifestyles, the compilers often create “models” to identify those likely to have similar traits. We have found this to be extraordinarily inaccurate in many cases and should be considered carefully.
When buying response type data from a website or reseller you’ll see the select but will never know whether it is all self-reported (with all its pitfall’s) or includes modeled data.As an example, Epsilon/Equifax has 5,000 dog owners in a town of 75,000 and InfoGroup (Donnelley) showed well over 25,000. They “model” or “infer” that because 51+% of households in a given zip code with certain demographics (single family homeowners within certain ages and incomes with children present, etc.) has a dog, all do. This is clearly untrue as some prefer cats, yet we find it typical of InfoGroup and other compilers’ data.
4. In addition to survey and product registration generated lists, seminar and trade show attendee lists and web-site inquirers are also widely available. Here again, there are many unqualified individuals in these lists and they must be carefully scrutinized.
Those attending business trade shows are often salespeople that sell goods similar to those offered by the booth vendors and have no purchasing power (or interest). Similarly, those attending consumer oriented trade shows (home shows, computer shows, auto shows, etc.), are not necessarily buyers and are just looking. Web site inquirers are especially suspect as many register false names and addresses. Obviously, “buyers” tend to be far more accurate than “inquirers” and “attendees”.