The latest news from the meaning blog

 

DataDynamic reviewed

In Brief

What it does

Desktop or web-based tabulation and charting tools for researchers or end-users with an integrated script-based data-processing module for data specialists. It can also be used to build data portals and dashboard reporting systems.

Supplier

Intellex Dynamic Reporting

Our ratings

Score 3.5 out of 5Ease of use

Score 4 out of 5Compatibility with other software

Score 4.5 out of 5Value for money

Cost

In euro (€): Offline tool €1100 per user per annum. Online: €1500 set-up fee, €1800 annual fee, plus €225 charge per person and per project.

Pros

  • Easy import from SPSS .sav files or Triple-S
  • Extendable gallery of output styles for both tabs and graphs
  • Powerful editing and data preparation workflow
  • Advanced machine-learning-based coding module for verbatim responses

Cons

  • Restricted filtering within online and desktop tools
  • Very limited range of statistics
  • Some limits on dynamic links to PowerPoint
  • No specific support for multi-language studies; interface is English only

In Depth

DataDynamic is a new arrival on the MR software stage. But is there space for yet another tab program? And is there anything significantly different about this one? Actually, there is on both fronts. While a majority of MR analysis tools effectively take their cue from the Quantum/Quanvert model, with the online tool existing as an add-on stage to a data processing activity, and the end-user working on a closed database of results, DataDynamic takes the more open SPSS as its muse.

It is often overlooked just how many researchers around the world use SPSS to do all their analysis on their quantitative surveys. On the whole, SPSS does a decent job for the market researcher wanting to analyse their own data, but it has its downsides. There is a steepish learning curve and the problem of picking your way though a host of options that are either rarely or never used. It is also a struggle to produce report-ready output for Word reports or PowerPoint briefings and summaries. But SPSS does make the raw data readily available to the researcher for them to work on and even edit. As a means of distributing data to other users this can also be a liability.

DataDynamic can work as a desktop tool, like SPSS, or as an online, web browser-based tool like SPSS MR Tables or Confirmit’s Reportal. The desktop tool (but not the online tool) also carries with it a complete interactive suite for coding and editing your data, which is aimed at the researcher as much as the data-processing specialist. It gives this product appeal to those researchers who are or like to be self-contained in their data processing capabilities. Better still, it provides them with the means to publish results to clients in a variety of ways appropriate to the needs of different audiences for research data.

For those that just want a dashboard with a few KPIs every month, there is a disarmingly simple process to create and publish these as web reports for clients to access securely in a data portal. These offer a single scrolling page of side-by-side tables, charts and commentaries, which can then be refreshed automatically as each subsequent wave of data is added. Users can look and copy data into their own reports, but cannot change it.

For those who want to dig into the data, their online log-on can be made to unlock access to the online cross-tab and charting capabilities. These are more or less identical to the capabilities of the desktop cross-tab and charting tool which is the core of DataDynamic. In either version it uses a familiar drag and drop technique to allow you to build cross-tabs from a structured list of questions. It is quick to put together tables, and there are all the options you would expect there to vary percentages, rank answers in order and apply or remove filtering, weighting or presentation options. Strangely, it only offers one significance test at present – the greatly abused t-test, which can risk being a safety blanket lined with asbestos.

On the other hand it contains a marvellous tool for creating your own target groups or profiles, such as those derived from segmentation models (though you would need to use something else to produce segmentations). It also scores two out of three for weighting: you can apply respondent weighting, and you can apply projections to a population total. However, you can only calculate simple arithmetic weights – there is no iterative model for creating so-called rim or target weights. Filtering too, is an oddity. It is quick and easy to apply any answer as a filter and combine answers from the same question, but it assumes you would always want to ‘or’ answers from the same question or ‘and’ answers from different questions. And if you create a target group, you cannot apply this as a filter – which would be handy.

Several of these restrictions can be overcome by using scripting. A powerful hidden feature of DataDynamic is the Visual Basic scripts that drive it. End users are unaware that these are being created as they compose their tables and charts, but they can be captured and edited, or folded into larger scripts to automate report production. It is akin to syntax in SPSS.

Other strong points include clear, attractive charting capabilities, based on the Microsoft Office charting engine, with user-definable template galleries; a surprisingly sophisticated suite for coding data, which even includes a trainable coding engine which will then automatically code similar datasets on the basis of examples you have provided before, and a great range of selective and cumulative imports from either SPSS or Triple-S data.

There are several other areas where more depth of functionality is needed. There is currently no real support for presenting or publishing results in more than one language for users to select their preferred language, for instance, and there some difficulties in publishing dynamic reports with charts that will refresh automatically, due to oddities in the Microsoft charting engine.

What I find most tantalising is that Intellex have used this platform to build a number of bespoke enterprise dashboard and drill-down reporting systems. At the moment, DataDynamic as shipped does not have all the tools needed to create your own enterprise feedback system, particularly in the area of user and data permissions. But enterprise reporting is something Intellex are planning to develop further and, if so, they could possibly be the first to market with a dashboard or EFM product that will work with research data from any source.

Customer Viewpoint: Yumi Stamet, Intelligence Group, Rotterdam

Intelligence Group is a research and consultancy firm based in Rotterdam in the Netherlands, specialising in the employment and recruitment research. It publishes, quarterly, a rolling two-year survey of the Dutch labour market comprising some 32,000 interviews to a wide range of commercial and public sector clients. It is a substantial survey with a large number of variables which it now distributes very effectively using the offline version of DataDynamic. Yumi Stamet, Operations Manager, explains: “We have been providing the data to the customers for a number of years using another software package, but we were not very happy with the way this software forced us to work, so we wanted to find a new way to process and deliver our data. The software had to be quicker and better – what we were using before was very slow – and of course, it had to be more user friendly”.

An important obstacle to overcome was the processing of the data, prior to it being ready for analysis and distribution, including coding a large amount of unstructured data and also applying weighting to balance the data.

“Intellex were very helpful in brainstorming on how it could be made easier and faster,” Yumi continues. “When we got the data into DataDynamic, they were able to help us to automate the process with scripts.”

As the new data was released using DataDynamic, Yumi was concerned whether clients familiar with the previous software would warm to it. “When they saw it they were very enthusiastic, particularly because they were now able to run a complete analysis of the dataset in a matter of minutes. It is a lot faster, and it looks a lot better too. Our clients also liked being able to make their own charts – and many other things are easier too, like sorting items, and above everything else the ability to create target groups very easily. Creating a target group just takes a few clicks, and it is easy to go back and refine it if it is not exactly what you want. A big advantage to me is that the software is very stable. With other software programs they can tend to crash if you ask to much of them, but you seldom see that with DataDynamic.”

To any company considering using DataDynamic, she advises: “It is very important you have at least one person learn how to do scripting as it will save you a lot of time. You don’t need to be a programmer, but it should be someone familiar with the dataset and somewhat IT-minded. If you look at all the time we have saved – with the preparation, I think we have gone from almost 24 hours to 12, and most of that time is because of the coding we have to do. With producing the reports, we have gone from a day’s work to just a couple hours.”

A version of this review first appeared in Research, the magazine of the Market Research Society, February 2008, Issue 500

Converso Enterprise reviewed

In Brief

What it does

Platform-independent Java-based multi-modal interviewing and analysis platform with an integrated portal-style front-end

Supplier

Conversoft, France

Our ratings

Score 3.5 out of 5Ease of use

Score 4 out of 5Compatibility with other software

Score 3 out of 5Value for money

Cost

In euro (€): Most modules €3,000 per user, plus €20,000 for Enterprise platform and €6,000 for web server module: all one-off costs. Maintenance: 18% of licence cost annually, or 25% for ‘gold’ support.

Pros

  • Extremely easy to use for moderators and participants
  • Can present a wide range of stimulus material
  • Offers several novel research techniques
  • Provides a complete transcript for analysis at the end

Cons

  • Only supports Windows for both moderator and participants
  • Not completely DIY yet: management module to be developed
  • Real time groups only – no support for asynchronous participation

In Depth

Converso Enterprise is an ambitious redevelopment project which deserves much praise for embracing what Web 2.0 technology has to offer head on. The portal-building and alert capabilities are excellent, and the main data collection platform is robust and sophisticated. But as an end-to-end solution it is still very much a work in progress. Substantial chunks are ready for production work now, but the gaps within and between these modules are just a bit too wide for comfort at this point. Given Conversoft’s recent rate of development, It is likely to look much more complete in as little as six month’s time, so for anyone planning to upscale their software platform next year, this is definitely one for the short list.

Converso Enterprise follows an entirely different architectural principal to most of the other new-generation research platforms on the market. Conversoft rejected developing in Microsoft’s .NET framework in favour of using Java, both J2EE, for desktops, laptops and servers, and J2ME (the flavour of Java for mobile devices). This approach does not on its own give the product a Web 2.0 pedigree, but it is a good start. It means that the software is totally platform-independent, so all users – researchers, respondents, technicians or end-clients – can use the browser or the operating system they want – Apple, Linux or any of the Windows varieties. This technical agnosticism extends to the relational database at the heart of the product, for survey data and panels, if used, which could be any of the modern database platforms – Microsoft, Oracle, or open source databases like MySQL or Postgres.

Conversoft also intends to create an open-source development platform to allow customers to extend the capabilities of Converso for themselves, but this does not exist yet.

What does exist is a wonderful portal-building tool that lets you snap into place any of the components of the Enterprise toolkit. You can create your own portal just for you or for entire groups of users – and then you can selectively switch on controls that will allow them to tailor the portal you gave them, to add in their own favourite things.

It could be the survey editing tool, a summary report showing the latest set of KPIs, an RSS news feed from the BBC or a link to Google Maps. This is where it gets exciting, because, once the missing developer tools have been developed, the techie people would be able to build whatever components you wanted to create so called ‘mash-ups’ of data from different sources on the internet, alongside your survey data – for example, to present geodemographic data in map form. What is more, Converso Enterprise components can be used as applets in other portals – so you could broadcast your poll results to other sites, or even Facebook.

Already, there is a rich library of components to choose from, particularly in the reporting area – which was never a strength for Converso in the past. It is relatively straightforward to create client data portals and dashboards that will present data graphically or as cross-tabs, or use intelligent reporting methods to highlight exceptions and provide alerts. Alerts are defined as triggers – really rather like dynamic filters that operate against the data and present a message. It all works fine with published data, but at the moment, you would struggle to show any real-time data from live surveys – such as to track response or get a live snapshot in a topline report. For these you need to resort to some of the legacy modules still.

Similarly, you can deploy new surveys through the portal, define your sample, and even use the very comprehensive access rights management tool from the portal – all of these are java programs. But the survey authoring tool is still a Windows program, and uses the old and rather complicated Converso scripting interface. That is promised for later this year, although nothing was available for Interface to obtain a preview.

These are not the only gaps waiting to be plugged. These are being addressed – and they need to be – though given Conversoft’s recent track record, the current feeling of being on a new highway where the cones are still in place, should have gone by the middle of 2008.

On the plus side, there is true multimodal interviewing with CATI. Web CATI is an integrated and very versatile handheld interviewing capability that will work on a very broad range of smartphones and BlackBerrys. The mobile interviewing is a new development and is impressive. There seems to be complete backwards compatibility with the old Windows-based CATI too.

Panel management exists but is not fully developed yet – the main panel management and respondent selection capabilities are there, but the panellist recruitment and community part is still missing. When it comes, it will offer integration with CRM systems, to use customers as a sample source, or to create customer panels.

The analytical tools are starting to look impressive too. A range of tables and charts can be created and presented directly in Word, Excel or Powerpoint, and it will populate native Excel or PowerPoint objects with data, to permit dynamic linkage. But if you wish to move data out to other MR analysis tools, then you are stuck until the planned Dimensions and Triple-S links are ready.

Converso Enterprise is an ambitious redevelopment project which deserves much praise for embracing what Web 2.0 technology has to offer. The portal-building and alert capabilities are excellent, and the main data collection platform is robust and sophisticated. But as an end-to-end solution it is still very much a work in progress. Substantial chunks are ready for production work now, but the gaps within and between these modules are just a bit too wide for comfort at this point. Given Conversoft’s recent rate of development, it is likely to look much more complete in as little as six months’ time, so for anyone planning to upscale their software platform next year, this is definitely one for the shortlist.

A version of this review first appeared in Research, the magazine of the Market Research Society, November 2007, Issue 498