Improving Agricultural Market Performance:
Developing Agricultural Market Information Systems
20
are often sold by non-standardised volumes rather than by weight) and prices can vary
significantly in a single day in the same market (FAO, 2017). Many of the methods used to collect
the prices are still traditional, that involves paper-based market surveys before the data is
transferred to a computer, a process which has been identified as a source of human error
(Galtier et al., 2014). In addition, reported prices often vary from actual prices for various
reasons, including exaggerated reporting, under-reporting of ‘extra’ values that come from the
buyer's relationship with the seller (e.g., credit, extra ‘gifting’ to regular customers etc).
Figure 3: Iterative Price Data Collection Process for MIS
Source: Adapted from CTA, (2015b)/FAO, (2017)In contrast, digital data collection is more reliable and economical; enabling processing and
dissemination of data to be much faster compared to traditional forms of data collection. Using
a variety of sources is also recommended for MIS (primary and secondary which depend on the
purpose of the MIS (i.e. clientele and what information is being provided). For example, primary
data collection through observation or survey and the use of secondary sources such as
government statistics/FAO statistics. Online databases providing application programming
interfaces (APIs) with which third parties extract data using programming provides an
opportunity for MIS to aggregate data from a variety of online sources on a continuous basis
with little or no manual work required (CTA, 2015).
• Select products related to
clients’ interests
• Consider where to collect
data from i.e. type of
market (e.g. retail,
wholesale, local) and how
many.
• Characteristics of the
products e.g. quality/grade
• Units of measurements
(local metrics and then
convert to kilograms if
necessary)
• Decide on frequency of
sampling e.g. monthly,
weekly seasonal.
• Decide on number of
observations (sampling
design and time of survey).
• Who will collect the data and
how (enumerators, paper
based or screen based)
• Design form train enumerators
pilot/ collect data
• Maintain, consult/get feedback,
support and upgrade