TMC lab reviews. (TMC Labs).

matchit v4.1

helpIT systems ltd.

1670 S. Amphlett Blvd., Ste. 214

San Mateo, CA 99402-2511

Ph: 866-matchIT (866-628-2448)

Web: www.helpit.com

Price: $1,495 for matchIT Lite, $2,995 for matchIT Plus, $4,995 for matchIT Pro, $5,995 for matchIT Campaign, $995 for matchIT Contact

In any teleservices agency. call center or other business with large customer databases, having duplicate records results in wasted mailings, extra postage, higher phone bills due to repeat phone calls and other costs. In fact, duplicate records can cost companies hundreds of thousands, if not millions of dollars per year. The technology to find and remove duplicates as well as prevent duplicate data during data entry has improved greatly over the past 10 years.

In fact, helpIT systems, a company headquartered in the U.K., has created a product suite called matchIT v4.1 that is extremely adept at finding and removing duplicates, and can even handle typos and phonetic variations of the same contact. It consists of several optional modules, some of which are specific to the U.K.., such as address correction and mailing/pre-sorting. However, helpIT has 'Americanized" the main matchIT application, which features deduplication, merge/purge, salutations and casing and job automation. We focused our testing on this Americanized version.

We should also mention that they offer another product called findIT which is an ActiveX control that can prevent duplicates as data are being keyed in. Another interesting product sold by helpIT is called marchIT Contact, which is used to merge data from duplicate records within GoldMine. We tested this product as well,

Operational Testing 

We started by testing matchIT v4.1, We launched marchIT and were greeted with a user-friendly splash screen with three options that helped get the ball rolling, including "Use The Setup Wizard," "Start A Job Script" and "Close This Screen." We clicked on "Use The Setup Wizard," which then prompted us to open a database file. We opened one of the sample database files which contained some duplicate records. A prompt asked us, "Do you want the Setup Wizard to determine the type of data in each field?" This indicated to us that the program automatically figures out which field is the contact name, which field is the address, etc. We let the program automatically figure out the field, names (see Figure 1) and noticed it did a pretty good job at determining the fields, After defining the fields, we clicked on "continue," defined our DBF file to import into and then clicked on "Find Matches." A report was displayed on the screen that contained a summary of the data imported. After closing this report, marchIT then displayed the "Internal Matching Summary" report, which reports. on the number of duplicates found by specific match key and by the score range (see Figure 2).

Next, we were given the option of viewing matches, verifying matches or deleting matches. We clicked on" Verify Matches" to see just how well matchIT discovered duplicates. In Figure 3, fields in red are different, yellow shows fields which are contained in the same field in the other record and white displays those fields which are identical, Even though most of the fields are marked in red (different), matchIT was smart enough to realize that "Mr. W. Dayton" and "Bill Deighton, Jr." are the same person. If you take our the title (Mr.) you are left with "W. Dayton," which could be "William Dayton." "Dayton" sounds phonetically very similar to "Deighton" and marchIT understood that "Bill" is a short form of "William." By examining the other fields and finding more similarities, matchIT is able to increase the score accordingly. The higher the score, the more likely the match, It can even match by common accents, such as someone from Boston pronouncing the name "Shaw" as "Shore."

If we see a record tagged as a duplicate incorrectly, we can click the False Match" button. Rather than verify each duplicate one at a time, if we noticed that at a particular score it is performing 100 percent accurate matching, we can simply continue to the next phase and specify this threshold score for purging duplicate records. matchIT can even identify which of a pair of records is "best" and should be preserved, or combine data from duplicate records so nothing is lost.

In addition to deduping a single file, the product also allows you to view "common ground:' or overlap of duplicates across two files. We were able to run a report that just gives a statistical summary of the amount of overlap between two files as well as runs a more detailed report that actually displays the overlapping records.

Next, we tested matchIT Contact on our live GoldMine database. matchIT contact essentially uses the same matchIT deduping engine, but we weren't able to customize any of the settings. We simply had to open GoldMine and then launch either matchIT's "Dedupe By Company" or "Dedupe By Individual." Since we have multiple contacts in our database at one company, we ran the "Dedupe By Individual" program. The report indicated that we had 977 duplicates in our database, We were then able to manually inspect each of the duplicates and determine if it was indeed a duplicate. As we browsed through each duplicate, the score at the bottom would progressively increase, We did notice some false matches at the lowest score (80), however once we reached a score of 107, we felt comfortable that there were no false positives. Thus, we set the "Threshold score" to 107 and let the software mark the "Ext3" field in GoldMine with a unique "matchIT" code so we could then use GoldMines Merge/Purge Wizard to merge the duplicate recor ds.

Features

matchIT will perform the following functions:

* Take care of casing, determine gender for creating salutations,

* Set up automatic job scripts for tasks you will repeat on a regular basis,

* Perform deduplication of any data; business or consumer and U.S. or international,

* Case data intelligently and create salutations,

* Correct and enhance addresses and verify, correct or add postcodes (U.K. only currently),

* Merge files, purge customers from lists, transfer valuable information between marching records (U.K. only currently),

* Suppress against industry-standard files such as MPS, GAS, TBR and NSF,

* Allow users access to a "drag-and-drop" Job Manager, which allows users to set up complex multifile jobs via one graphical form,

* Intelligently merge data when deleting, transferring, merging or verifying dupes, to create a "meta record" containing all the best quality data, e.g., phone numbers, product codes, name elements etc.,

* Use ODBC for Access, SQL Server, DB2 and Oracle, on both Import and Output, and

* Offer a multiuser network version, allowing a number of concurrent users.

Room For Improvement

The GoldMine-based marchIT Contact needs more user control of the matching scoring. We found that many of the lowerscoring dupes (score 80) were duplicates; however, some were nor. It was relying too heavily on the ZIP code. As a result, it would match if there were two common names at two different companies but within the same ZIP code. For example, there could be a John Smith at IBM and a John Smith at ACME Corp., both within the 06483 ZIP code, and matchIT Contact would score this as an 80. We wished we could more heavily weight the company name and address, so it wouldn't score this as a dupe. Unfortunately, unlike matchIT, matchIT Contact doesn't allow us to customize the weighting. We did contact helpIT and they told us they could customize this application at a customer s request. One final note regarding GoldMine support -- helpIT informed us they have a beta version of data-entry plug-in for GoldMine that will prevent duplicate records at the time the record is being added,

Conclusion

helpIT claims that matchIT uses the most "fuzzy matching" algorithms on the market. We were impressed with its ability to find phonetic and miskeyed matches, most of which would usually remain undetected, marchIT matches by comparing the data in great detail, without relying on any single item of data being consistent,

Each time you send a duplicate piece of mail, it costs you money and annoys recipients. matchIT v4.1 is very easy to use. It's an intelligent deduping software package that does what it's supposed to do and it does it very well. Ultimately, we were also impressed with matchIT Contact, which quickly and easily tagged duplicate GoldMine records for merging. Overall, 7MG Labs was quite impressed with the matchIT product suite and would highly recommend it to anyone with small, medium or large databases.

RATINGS (0-5)

Installation: 4.5

Documentation: 4.75

Features: 5

GUI: 4,75

Overall: A

For information and subscriptions, visit www.TMCnet.com or call 203-852-6800.