How You Can Use Evaluation Data To Fine-Tune Your
by Bill Goodwill Chairman, Partners in Public
National Broadcast Association for Community Affairs
Houston, TX 1
Developing a comprehensive evaluation system for your
PSA program is critical to measuring your success and
being able to justify the worth of your efforts. In
the old days getting reliable evaluation data was
difficult if not impossible. However, there are now at
least two major electronic tracking systems for TV and
even a system that can track radio in major markets
that provide very reliable evaluation feedback.
Since broadcast television accounts for the
lion’s share of PSA exposure, we spend more time
tracking and analyzing TV data than that resulting from
other forms of media exposure. Following are just a
few of the key trends we measure for client campaigns,
most of which rely on electronic data provided by A.C.
Nielsen’ SIGMA electronic tracking service.
One of the most frequently asked questions
posed by clients is: “How is my campaign doing….?”
And of course, we are thinking….”compared to what….”
since it is a relative question. To provide a meaningful response,
we create campaign benchmarks for each media type that permit us to
compare any given campaign against a standard. Not only does it give
us as evaluators and our clients some basis of comparison; it also allows
us to watch trends develop over time and apply corrective actions when
we see a campaign performing significantly under the benchmark.
The data in the foregoing graph also provides
meaningful feedback for future budgeting purposes
also. While it is true that radio and print are going
to help you reach discrete audiences, the benchmark
clearly shows that broadcast television is going to
generate the lion’s share of reach for your
One of the most frequently repeated mis-statements about PSAs is that
"they are only aired during junk time"' which is usually referred to
as very late in the evening and early AM. As this graph shows, that
is clearly not the case. The electronic data provided by the two major
monitoring services breaks down PSA usage by different times within
a 24 hour period and this campaign nearly two thirds of PSAs air in
the best dayparts, which start with the Early AM daypart up to the first
half of the Late Evening daypart (10PM-midnight).
Depending upon your particular target audience, you
can track the percentage of PSAs aired during the times when your target
audience is most likely viewing TV, which provides an important quality
indicator. For example, women (yes there are still stay-at-home moms
- just ask Proctor & Gamble) are heavy TV watchers during the day.
Spot Length Analysis
Perhaps the second most widely held misperception about
PSAs is that stations use mostly shorter spot lengths.
Again, the data shows just the opposite.
Referring to our benchmark again, SIGMA data from
campaigns which included sixties in their package,
shows that station using sixties ranged from 56% on the
low end, to 61% on the high end, with an average of
57%. This graph has rather obvious implications to your
creative team. If they are producing only shorter
length spots, you are missing an opportunity to
register a more compelling message that is possible
through longer lengths. Perhaps more importantly, a
:60 is worth twice a :30 and four times a :15, meaning
you will generate much greater value for your campaign,
to many people the most important determinant of
Finally, if you are using a tag with a website or an
800 number in your PSAs, you need the longer length
messages to adequately register the information on the
Analysis By Market Size
Another way to look at PSA data is to analyze usage by
market size. Nielsen uses DMAs - Designated Market
Areas which range from 1 (New York being the largest
population center) to 212 the least populated market.
By examining the percentage of your PSAs that are being
used in various markets, which are correlated to
population, you can determine areas where your exposure
is strongest or weakest. This data provides feedback
for taking corrective action in markets where your
exposure is weak or nonexistent. In the case of the
Peace Corps campaign, nearly two-thirds of all exposure
occurred in the top 100 markets which comprise about
80% of the U.S. Population.
Electronic monitoring data - specifically Nielsen's
Post-buy package, which works in tandem with SIGMA, can
permit you to measure impact upon a specific target
audience. By examining demographics, you can track the
number of gross impressions delivered to these
Usage By Type
It is important to know as much as you can about the type of TV outlets
that are using your PSAs, and what they contribute to your total exposure.
You can analyze usage by media type, i.e, broadcast versus cable TV,
radio or by specific types of media as shown in the next graph.
You should know specifically the type of stations using your PSAs,
so you can fine-tune future distribution strategy. For example, are
they network affiliates…. independents... cable networks....or
individual cable systems? What are the formats of the radio stations
using your PSAs and are they the ones that reach your primary audiences?
Perhaps you are interested in reach a youth audience, in which case
you want to examine the number of college stations using your PSAs,
or perhaps audiences by ethnicity such as Hispanic, African-American
or other minorities. By analyzing the level of usage among different
types of stations, you can determine if you reached your objectives.
Level Of Response
Perhaps the most widely held misperception about PSAs
is that "they just don't work... if you want to get
exposure for your issue, you have to buy the time."
Once again, TV PSA data paints a different picture.
Unknown to many PSA planners, PSA campaigns can even
generate excellent levels of response. For years the
primary method the Peace Corps used to communicate its
message to young prospective volunteers was a national
PSA campaign. They carefully tracked the source of all
incoming leads and applications and shown here they
generated 782,000 leads during the six year period we
worked with them, resulting in 58,588 applications.
The foregoing graph also shows another interesting
trend demonstrating the efficacy of PSA campaigns. In
the first five years, they distributed two campaigns a
year which are the years when leads were on the rise.
In 1998-99, they scaled back to only a single campaign
and experienced a reduction in both leads and
One of the most definitive studies showing the impact
of a toll-free number used in TV PSAs was done by the
National Cancer Institute (NCI) which measured calls to
its Cancer Information Service.
NCI distributes public information via a number of
different media and was trying to determine the impact
of TV PSAs in general and the affect they had on phone
calls to its toll-free number. In the months prior to
the release of the TV PSA, phone calls averaged 1,872
per month as shown by the flat yellow line for
July-November. After the TV PSAs were distributed,
over 71,000 calls were logged during the eight month
test period, or an average of 8,951 calls per month.
Also, as the graph shows, there was a direct
correlation between calls and level of PSA activity.
By using a dedicated 800 line and tagging your PSAs
with the phone number or a website URL that is
specifically set up to track “hits”
resulting from your campaign, you can get excellent
feedback on PSA effectiveness, as well as demographic
data on typical responders. Or, you can ask for people
to write for specific pieces of literature that have
been prepared exclusively for use with your PSA
campaign and track the number of requests you receive.
PSA usage data can be broken out to demonstrate where
you are getting the most or least exposure by field
office, chapter, district or some other geographic
configuration as shown in these graphs.
Obviously there are parts of the country where there are more media
outlets than others, or you may have distributed more
PSAs to certain markets or regions. However, by tracking usage by some
discrete geographic method, interesting patterns begin to emerge and
by sharing this trend data with your community partners, hopefully proactive
tactics can be employed to reverse negative patterns. In these two scenarios,
you can graph the number of PSAs distributed versus the amount of exposure
you get which tells an interesting story and provides the basis for
By sharing this type of data with your community
partners, it helps involve them in your campaign
strategy, and shows specifically where they need to
concentrate future media contacts.
Applying the Previous Usage Index
On each of our client reporting portal websites we provide distribution
lists indicating which stations have or have not used a client’s
PSAs in the past and the Previous Usage Index (PUI) for that station.
By looking at these two columns as shown here, we can formulate a follow-up
strategy very easily. For example, in this case, KDMD has not used
this client’s PSAs, yet they have used our other clients’
PSAs 50 times in the past.
Another way to look at users versus non-users is shown in this graph.
For a Social Security TV PSA campaign, there were 46 stations that did
not use SSA PSA but have used other client PSAs from one to more than
25 times. These are examples of looking at "actionable" evaluation
data that can make a very significant difference in the ultimate usage
and impact of any PSA campaign.
Confirmation Of Shelf-Life
Before electronic monitoring data was available,
everyone estimated the shelf life of PSA campaigns to
be about 12 weeks, because that's when bounce-back
cards began to dwindle.
Now, through electronic monitoring, we know that
campaigns still get fairly good exposure at 26 weeks,
52 weeks and even up to a year and a half after they
are released. As shown in this graph, a campaign for
Volunteers of America generated 55% more value at the
end of two years compared to the value after only
tracking for one year.
An Army National Guard TV PSA campaign was generating
significant levels of exposure 17 months after release.
This is useful intelligence because if your campaign is
still doing very well at the time when tracking is
about to expire, you may want to consider extending
tracking for a longer period.
Perhaps the most important reason for evaluating any
campaign is to substantiate the overall effectiveness
of your program. If you can't prove the effectiveness
of your program to your boss, or whoever else is
funding your campaign, it probably won't survive in
these lean times.
There are lots of ways to do this:
- Produce a monthly management summary of the dollar values
generated, the number of PSAs aired, and the number of stations and
markets where your PSAs were used, the types of audiences reached
and gross impressions. (We do this for all our client campaigns.)
- Circulate the summary to all stakeholders in your organization who
has any interest in your brand image... board members, volunteer leaders,
senior management, field public affairs personnel and funding sources.
you use an 800 number in your PSAs, you should be able to track the
number of phone calls generated, website visits or literature requests
received. All of these organizational goals or campaign objectives.
- Compare key trends data from one campaign to another, from one
year to another to determine whether you are making progress over
- Calculate the advertising equivalency value – the value you
would have spent if you bought the airtime - and compute a cost-benefit
ratio by dividing the production and distribution costs for your campaign
by the amount of exposure received.
Keep in mind that there are limits to corrective
actions that you can take at the national level since
real change takes place out in the countless
communities across our great land. However, by sharing
national campaign strategy with your community
partners, analyzing the data carefully and sharing
results with your field staff, you can generate the
kind of support you need to reach your goals.
And most importantly, remember that evaluation data should be actionable.
Compiling evaluation data just for the sake of having numbers is a meaningless
exercise if you don't use the data to change the way you develop future
1 Article has been
updated with later evaluation data and trends.