Dapresy Pro 4.3 reviewed

Flick to the end of any research brief presented to an MR agency these days and the deliverables section is very likely to speak of “secure online access to the data”, “dashboard-style reporting” or “real-time delivery of results to stakeholders”. These short phrases can be a snare to the unwary – a tripwire to a landmine that can blow a significant hole into budget and resources, and leave account managers and technicians feeling battered and bloodied by the experience of trying to deliver them on a tight budget.

The difficulty is twofold. First, the requirements can vary enormously, and tend to escalate into something more complex and technically demanding than was expected when the project was costed. Second, the tools the technicians are often forced to use to custom-build these systems from scratch (like Xcelsius, Dundas Charts and Microsoft Sharepoint) offer very little in the way of support for the kinds of data or analysis we take for granted in market – which also sends time and costs spiralling.

This is specifically the problem that Dapresy, a Swedish software developer, set out to solve a few years ago by developing an environment that would let you create web-based data reporting systems that assumed there would be a mix of survey data and other data to present. The firm has developed Dapresy Pro on the back of custom reporting systems it had developed for research companies in the Nordic region. Dapresy aims to turn portal creation for dashboards and complex stakeholder reports into a point-and-click operation.

The product is marketed as four different dedicated applications, for trackers, stakeholder reports (i.e. a replicated set of reports for many different users, each only showing data relevant to each user), mystery shopping and employee research. This means there are canned reports and templates available for each of these kinds of activities, but underneath this there are in fact two different software modules within the Dapresy Pro platform: the tracking module and the stakeholder module which streamline the creation and delivery process.

It is a virtue of the product that it covers a very broad spectrum of the kinds of reporting that research buyers are calling for, from brand trackers through to complex hierarchical ad hoc or dipstick studies. My estimate is that it probably does 80% of what gets asked for – which is quite an achievement. This could translate to 99% of what you need on some projects, although it will be less on others.

The basic process for creating and publishing is relatively straightforward. All of the set-up is done using a simple web-based interface. The most technical part is the data mapping stage, in which you import the data. Dapresy’s preferred format is SPSS, but they have integrated with others including the Dimensions Data Model using XML – and they are looking at Triple S.

It will also detect changes in subsequent imports, which is essential for trackers – but the detection is not foolproof and you need to be pretty vigilant if working on a fast-moving tracker to avoid error. Once nice feature is being able to import data and activate different waves or sections independently. It is only when you activate data that it is published to all portal users.

Beyond the import, there are separate areas that lead you through the rest of the design process: creating report objects and selecting variables; designing the portal, defining users and user access rights.

You start by designing reports in a folder structure, that lets you relate reports to one another. When published, this structure will determine different tabbed windows within your portal and also the drop-down and other navigational features – it makes it easy to create an intuitive navigational structure for your client that you can test as you go.

Dapresy provides a Chart Lab in which you can create a very wide range of visualisations. Again, my guesstimate is their canned charts probably meet 70% of what gets asked for out of the box, but you do have the option for custom charts to be added. You can also arrange these components side by side to create a compact dashboard feel to your portal. You are limited to four objects on any one screen, though, which denies you the kind of ‘small multiple’ presentation you often see on a corporate dashboard with a lot similar charts presented side by side for at-a-glance appraisal.

The software effectively solves many other MR-specific data publishing headaches. It is impeccable in concealing small bases and any inadvertent disclosure by letting you define an anonymity limit. Yet on mystery shopping you can drill down to read individual observations and can also access photos or videos uploaded by the mystery shopper for that shop.

There is very sophisticated support for time series, based on interview date – and charts will take into account the fieldwork periods, allowing data to be rolled and unrolled the way most researchers expect. Trend lines are also adjusted, so that any missing periods are shown as a discontinuity in the chart, and not a steep descent to the zero line.

Drill-down is also something you control explicitly – you can define what, if anything, the user sees when they click on chart elements or data points. In the tracking module, you can even open up full cross-tab analysis to some users, or lock down the drill-down capabilities to so they only go to other static reports.

Access rights are simply and effectively managed using an Excel spreadsheet for many users, but the system will also support corporate single sign-on, using a proxy server and an encryption key to authenticate users without them having to log on a second time.

However, not all features are offered in both stakeholder and tracker modules. The tracker module is where the fancier analysis options reside, and the stakeholder module is where sophistication over user access is concentrated. You can’t pick and mix, because both modules use different databases, so that can leave you with a conundrum for some projects that sprawl in both directions.

The price is not cheap, but what it does is make cost estimation highly predictable, because the main component of the cost is something every research brief typically does contain: the overall sample size.