The paper discusses web-based data collection in the broader content of one of the most challenging market-research-tasks, the volumetric forecast for new concepts.

Decisions about new concepts based solely on information gathered in a typical concept test run the risk of ignoring critical factors of success. The total size of the potential market (the interested universe), the expected awareness and distribution level, fixed and variable costs per unit and the repurchase rate after initial trial are crucial information for determining the economic success of a new product. A volumetric forecast estimates the total sales and the revenue in a certain time period after launch and incorporates both survey and external data.

A volumetric forecast faces two major challenges regardless of the data collection method:

1. Predicting the awareness level based on planned marketing actions

2. Ensuring the representativeness of the sample for the interested universe

The awareness level is mathematically modelled as a function of the reach of a marketing campaign: Information regarding reach is provided by media planning and is defined as the probability of exposure to marketing based on spending levels and targeting. The mathematical model takes into account reach from a variety of sources including in-store promotion, couponing and sampling.

Representativeness of the sample for the interested universe is obtained by propensity weighting. Propensity weighting controls for subtle differences between target population and members of an online access panel that do not result from obvious sociodemographic characteristics that are normally controlled in a quotation sample. In a combined representative CATI and panel-based online survey the probability of membership to the online access panel is estimated using a logistic regression incorporating a variety of questions about attitudes towards the internet. Afterwards the online sample is weighted according to the representative distribution of the propensity scores which indicate panel affinity.

The proposed paper describes the challenges and solutions in the volumetric forecast based on online access panels and demonstrates the general appropriateness of the method.