1) Given all the limitations of collecting data for this indicator (see the points below), it is recommended that you either use a different indicator or complement it with relatively more accurate indicators, such as Wealth Quintile or Household Consumption. Alternatively, focus on assessing the average income from a specific source only (for example, from the sale of agricultural production).
2) The vast majority of people – most likely including you – are uncomfortable about telling others about their real income. Respondents may evade questions or give inaccurate answers (due to various reasons, such as lack of trust, fear of being excluded from receiving external assistance, or taxation-related concerns). Before the data collector starts the interview, s/he needs to gain the respondent's trust – therefore, ensure that the data collectors explain carefully why your agency needs the data, how it will (not) be used and why it is important that the information the respondent provides is correct.
3) Household income is not necessarily a good indicator of household's well-being. A rural household that earns an average of 100 USD per month, does not pay any rent and produces most of its food might be much better off than an urban household earning 250 USD per month. Compare data from households with similar characteristics only.
4) The longer recall period you use, the less precise the answers you will get (people simply do not remember all the different income sources and the amounts they gained and very few of them keep any written records). At the same time, using a very short recall period call might capture only a part of the reality, resulting in receiving imprecise data. If people have a smaller number of more regular sources of income, you can use a longer recall period (e.g. 3 months). If it is the other way round, it is recommended to use a short recall period (e.g. one month).
5) Be aware that if you interview only one household member, s/he might not have a complete overview of the other members' income. If your survey's piloting shows such a situation, you might need to invest time in interviewing every earner in a household.
6) Income is often prone to significant seasonal variations (especially in the rural areas where people's income is determined by the sale of their harvest, seasonal employment, etc.). Do your best to collect baseline and endline data in the same period of a year; otherwise it is very likely that they will not be comparable. Furthermore, in many contexts, household income is very irregular – a household can earn very little for, let's say, two months and then receive a large one-off income in the following month. Before you conduct your survey, assess the income regularity (and seasonality) and adjust the survey's methodology accordingly.
7) Keep in mind that household income is not always distributed equally between members, resulting in gaining skewed data on the household members' access to income – especially with regards to women and children. Therefore, when measuring households' income, always complement it with a measurement of who decides how household income will be used.