Editor’s note: Tim Macer, managing director of U.K. consulting firm meaning ltd., writes as an independent software analyst and advisor.

You may be skeptical about running your office software under one of the Web-based ASP or “Software as a Service” (SaaS) solutions, and you may not be ready to move your financial accounting systems to SaaS quite yet. If the Internet is sluggish, it soon becomes tedious and unproductive to work on documents, spreadsheets and presentations - the screen is two steps behind you all the time. Enterprise-wide business-critical operations such as accounting open up issues of control and ownership, especially at key times such as tax reporting deadlines, when you may notice the servers taking a hammering from all the other customers.

Switching from a desktop-based survey analysis tool to a Web-based one, however, makes perfect sense. The odd bit of balkiness in the performance of the Web-based software is a price worth paying to let users get their hands on the data several days sooner or allow research departments to distribute the data to all their end-users with a few clicks. The issues of resource competition and ownership are more manageable too.

Yorba Linda, Calif., software firm Productive Access Inc. (PAI) perhaps took more of a gamble than other players, bearing in mind the name of the firm, when deciding to port its mTAB desktop survey analysis program to the Web, given the product’s 15-year pedigree. But the result, mTABweb, is a faithful reproduction of this ingenious-to-the-point-of-quirky analysis tool. It runs in a Web browser, with full drag-and-drop control, achieved through Java. It isn’t as fast as the desktop mTAB, but the Productive Access name is at least safe: If you have a stable Internet connection it gives perfectly acceptable performance. The software is optimized to run under Windows with Internet Explorer 6 or above but it also runs acceptably on other platforms, including on a Mac in Safari, in my tests (though slightly less successfully under Firefox).

Why quirky? Because the interface hinges - quite literally - on a simulated ring-bound desk organizer, with a series of tab dividers on either side that let you choose which two pages you wish to show side-by-side (Figure 1). You select the variables to tabulate from the Questions tab and by opening the Row tab or the Column tab you make your choices for stubs and banners. There are other tabs for choosing filters, switching datasets or adding in a third level beyond the banner and stub. Percentage and respondent base selections are easily selected from drop-down menus.

When your table is assembled, you click a button in the toolbar and the table is generated and presented in the same window, which now switches to a single panel. One of a line of buttons shows the ring-binder icon, and this lets you flip back to the set-up at any point. It looks disarmingly like an Excel spreadsheet, although it isn’t (Figure 2). However, it gives the output window a very intuitive feel to users. Buttons on the toolbar and a right-mouse button menu open up a wide range of options for finessing the output, from omitting columns or rows to selectively adding shading or borders.

Level of sophistication

The almost cheery simplicity of the interface and ease by which you can move from data to tables belies this program’s actual level of sophistication as a serious survey analysis tool. Look at any of the features or options and you will find an intelligent set of capabilities. If you need statistics, there are means, standard deviations, medians, chi-square and t- and z-test scores. It will also automatically create top-two and bottom-two box scores for any rating scale-type question without requiring any recoding.

Filters and new variables are also easily created using graphical editors. There is a range of built-in options for cutting numbers into categories or ranges, as well as intelligent handling of date fields. An interview date can be converted into a profiling variable based on calendar months or fiscal quarters with surprising ease, which is particularly handy on trackers.

The support for trackers goes much further. You can combine different datasets and there are tools for managing the differences between the waves of a tracker within the software.

Back in the analysis view, once you have viewed a table, you can save it, give it a name and come back to it later. You can also select a portion of it and turn it into a chart (Figure 3). There are a dozen chart styles to choose from, though the output styles are limited, compared to Excel or PowerPoint. However, you can also run correspondence analysis in the charting module and display these as maps. These too can be saved or pasted into presentations and reports.

Data available to many

PAI’s customers are typically research buyers rather than research companies. One such customer is Pulte Homes, a new-home builder operating across the United States. It uses mTAB and mTABweb to make survey data, such as information from its ongoing buyer’s profile survey, available to marketing managers and decision makers across the company.

Jim Rossiter is the firm’s vice president, strategic marketing for the Great Lakes and Northeast regions and has close to 15 years of experience using mTAB and mTABweb. “One of the great things about it is that by putting data into mTABweb it’s so easy to make that data available to many, many people around the company, instead of having to move files around or train other people on more complicated applications, and it is accessible to everyone. We can take a group of non-technical marketing analysts and put in their hands a lot of data they can manipulate very quickly.”

Rossiter is often involved in familiarizing new users to mTAB products, and finds that 30 minutes is usually sufficient to provide a useful grounding in the programs - enough for people to be able to dig into the data for themselves. “People are pretty impressed with how simple it is, especially people who do the same kinds of things in other applications and recognize how something that might take them 10 minutes in another application, they can do in 30 seconds in mTAB,” he says. “The whole power of the thing is the speed by which you can do these tasks. If it takes you too long to figure it out, you might never get to the results you want. If I have mTAB and can figure it out in minutes, I can probably come up with better answers than I would over the course of hours or days.”

Among his favorite features are those for recoding data and breaking the raw values into intervals and groups. “It’s fantastic to have a lot of detail in monetary values - in our database we will have the actual purchase price paid in specific dollars. I can very quickly look at customers that spend under this much one day, and then turn those into $10,000 ranges for something else I am looking at the next day. I always like to keep the detail and then aggregate up from that. And with dates, you can aggregate these into time periods and compare quarter-on-quarter - that is very useful.”

The ease by which findings can then be moved into Excel and PowerPoint is another advantage, he says. “Often the user will export data into an Excel form and generate presentation materials, which people are very comfortable doing. We also have some custom reports that we have set up with mTAB which export data into Excel on a quarterly basis and this is then posted up on a server where everyone has access to it.”

Direct import capability

One reason why the product has been less appealing to research companies is the business model behind the software: customers must send their data to PAI or one of its associated companies, which will convert the data into an mTAB project database. The software is essentially costed on a per-project basis, which includes this setup charge. However, this is likely to change when PAI introduces a direct import capability from SPSS datasets later this year. Other data-importing enhancements would open the software up considerably.

Compared to other Web-based end-user analysis tools, the other obvious missing element in the software is user rights management. While access can be restricted at a dataset level, large enterprises often need more granularity - such as to restrict certain variables or even turn on and off functionality within the software.

My only other negative observation is that the port to the Web, while it is faithful to the original, makes for a slightly strange-looking interface on the Internet and the program feel rather dated. Dare I even suggest it’s time to downplay the desk organizer look?

Overall, a better analogy for both mTAB and mTABweb is to a PDA such as the Palm or BlackBerry - it looks quite modest from the outside, but get it in your hands and you keep discovering more to use and more uses for it.