Summary of EmTAG Survey
The Emerging Technology Advisory Group (EmTAG) focuses on the emergence and impact of technologies on members of the Association of Imaging and Information Management (AIIM). EmTAG distributed surveys, meant to assess current trends and issues with technologies, at two sessions during the recent AIIM Conference.
Two Hundred Eighty Nine (289) surveys were returned and processed. A summary of the results indicate that the most important emerging technologies to impact business are: Document Management; Web Tools; Workflow; Image Management; High Speed Networks; Storage Technology; Groupware; and Object Oriented Development Tools. In addition, the major impediments to implementing new technologies include: Cost; Resistance to change; Lack of standards; Lack of top management buy in; and Too much reorganization. Additional impediments written several times in the survey results include the Lack of Resources/Staff and Unclear Requirements/Ownership.
Image management systems lead existing and planned implementation followed by Document Management, Workflow and Groupware. Workflow is showing the largest increase and earliest to be implemented compared to existing systems. Average cost of all systems was reported to be $3 to $5 million, while the median cost was given as $200,000 to $800,000.
Top areas within companies for implementing new technologies in document management include Finance and Accounting and Customer Service.
sSWAK - Survey Sealed With A Kiss
Several points of learning were made from putting together a fourteen question survey for the recent AIIM conference. One of the most interesting discoverys was a survey that was literally sealed with a bright red lip smack. It's difficult to determine exactly how much weight to give to this kiss but the form processing company swears that their competition made the entry to differentiate their products intelligent smooch recognition feature.
The EmTAG survey was doubled sided and contained fourteen multiple choice (select one or select many), and write-in the box questions. Optical Character Recognition and Mark Sense was used to analyze the results. The form was created with Microsoft Word with symbols for the fill-in circles and tables with borders for the write-in boxes. The form included black text with a colored (blue) border which was double-side printed on a highlight color printer. (Unfortunately the scanning was done edge to edge so the border added to the size of the images which increased the processing time.)
While the science of "Survey Research" was not strictly used, the survey results are most interesting. Survey Research can be used to determine the best way to represent the questions. (Several Universities are involved in research on this topic.) Both question order and answers within the question could have impacted the survey results. These factors were not considered as the survey population was not selected scientifically. Therefore, the absolute significance of the surveys can not be determined. The relevance of the results can be seen as 289 surveys were returned from a total of 635 attendees in two seminars. Of the returned surveys, 51 made additional comments and 53 asked for additional information on EmTAG (the committee that was making the presentations.)
The survey used the following types of instructions: "Please fill in circles and/or write in boxes", "Please complete questions on the back", "Choose best", "Choose all that apply", and "Please rank those that apply using numeral where 1 is the most…". A preliminary revision did not include several of the instructions and instead had the simple comment "You know what to do…." Of all audiences, it was thought that attendees of the AIIM conference would be superior form fillers. Unfortunately, several mistakes by respondents or poor assumptions by sponsors were made which affected automated processing.
Typical Mistakes
 | The following symbol is a fill circle. It was assumed that
people understood that the circle should be completely filled
in not just marked through. Several people drew a single line
through the circle.
| Several people marked the fill in boxes with an X instead
of the appropriate numeral.
| Some people used multiple 1's to change the weight of the
survey.
| Some people selected more than one choice when only one was
asked for.
| Several surveys had only one side filled in.
| |
The decision was made to allow the errors into the survey results instead of defining all the rules that would be required to automatically locate the errors. Correcting errors in the survey results seemed of less value than the information the respondent was trying to give.
These errors could cause problems in the survey results. Therefore, it is important to define error/default values for mistakes. These values could help provide support, but shouldn't be used to govern the results. Values were used that would be above the typical response; such as 4 on a "rank the top three" question.
It was very difficult to automatically process the results of the surveys. Few tools automatically analyze results of survey questions. Microsoft Excel and report tools provided in database systems can give some answers but this requires the generation of several queries for each choice which is a very time consuming process. The decision was made to write a general purpose tool in Visual Basic that used a Microsoft Access database. The results were even styled. Unfortunately, there was no simple way to compute complex trends so this will be left to future efforts.
Quick Smacks (Simple Recommendations)
| Make sure to pick up forms from the survey area. At least
25% of completed forms were not handed in, but were left in the
seat of the respondent.
| Complete instructions with examples are required. Not all
knowledgeable respondents will complete the form properly.
| If possible minimize the different types of questions so that
the respondent has a better chance of providing the requested
information.
| To increase the response rate, offer feedback for the respondents.
| Validation and results analysis have a cost. Determine what
value you have for precision and the survey data prior to the
survey.
| Make sure to clearly differentiate choices so that variation
in respondent interpretation can be minimized.
| Take care when processing double sided forms. Pages may be
jumbled in scanning which will make accurate analysis impossible.
It might be helpful to uniquely identify both sides of the form
so that data can be accurately recombined.
| |
Extended Smooch (Additional Consideration)
| Data validation improves accuracy. The importance of results
will indicate the value of validation. If data is to be used
for scientific or financial analysis then precision may be required
to be greater than 100%. One hundred percent accuracy for several
data items may be required, such as a check with numerals and
a written description of the amount. However, a reader response
card may not require accuracy in the request as much as in the
reader information section.
| Allow for additional comments to be made in the margins.
Whenever forms are created they won't address every need or be
understood by all respondents the same. Be sure to make allowance
for this. This mechanism can also be used to catch responses
that are corrected by crossing out an undesired result and rewriting
the true answer somewhere around the original area.
| Use write-in blanks to get extra values for the next survey.
If the same idea appears many times it should be taken seriously
as more people probably would have selected that choice it if
it were included. One write-in response of the survey had an
exact match by five people and three other people used abbreviations
that were close.
| |
Last Embrace (Conclusion)
In the long run, automatically processing this one form was significantly more sophisticated and took several more hours than a manual tally of the results would have taken. However, the results were loaded into a database and were available more quickly than would have been typically possible. This is yet another example of the need to weigh the benefits versus the costs of technology. We met our objectives of gathering and quickly reporting data on trends in emerging technology while using the technology that we are to be watching.
EmTAG is in the process of putting the current survey on the Web with running results. Look for it on a web site near you. EmTAG will also have another survey at next year's AIIM conference.
Clark Brady
Member of EmTAG - Emerging Technology Advisory Group of AIIM
Senior Scientific Analyst, Submission Systems, Eli Lilly and Company
I need to thank Grant Lindman for his editing assistance.
|
Last Modified: Aug 6,1996 Page Maintained by Clark Brady {cbrady@iquest.net} |