Implement TAM & SAM via Survey Methods
Using our typical survey methods, we can relatively easily calculate a TAM, and with some agreed-upon criteria with a client, a SAM. Generally speaking, we do not have the resources to calculate a SOM as this process requires a lot of secondary data from the client that will be outside of a survey we control. Using the SAM, we can provide various estimates of SOM based on the client’s estimated rate of market penetration.
We will now go into a generalized step-by-step process of creating a survey for a client that has the necessary components to estimate TAM and SAM.
First, we need to define the product/service that you will be measuring the TAM for. This will help you to determine the potential customer base.
At this stage, we need to ask the client what product or service they sell at a high level (e.g., clothing) and any specifics that may set their product or service apart from others (e.g., high-end, sustainable dresses). This may be a new product being added to their existing line, an add-on to an existing product or service, or a new market they are trying to enter with an existing product.
Example: If the client sells women’s clothing, we need to know if they are interested in a TAM for all of women’s clothing, or a specific product like dresses. Typically, the TAM will cover the larger category like women’s clothing, and the SAM will address the specific product like dresses. However, if the client is truly only interested in the TAM for dresses, then we should proceed accordingly. We will also need to know specifics about their clothing that sets them apart (e.g., Is it high-end, sustainable, or shipped on a recurring basis?). While knowing the general product will help us measure the TAM, all of the specifics will eventually help us shape a TAM into a SAM.
Next, we need to define the largest market that the client hopes to serve in the next 5 years or so (e.g., are they trying to be a national or global brand?) and which markets they currently to serve or will serve within a year. We should always be able to get census data from the defined market. We cannot conduct a TAM unless we can extrapolate the sample estimates the population. Typically, we will be running TAMs for the US market. In those cases, we can use the quotas set up in Trendlines.
Example: For our client who sells women’s clothing, we need to know the largest market they need to size for the TAM. Although a global TAM may be interesting, the data costs to get a representative population across the globe will be extremely high, and most of the data will be irrelevant if they do not plan on entering many of the markets. For many of our clients, the TAM will be national, or specific to a few countries. For the remainder of this example, let’s assume the client is interested in a TAM for the U.S. For the SAM, we will need to know the cities that they currently serve or plan on serving relatively soon.
Using the market definitions in step 2, we will need to create a sampling plan. For a TAM, we want the sample to be representative of the largest market the client hopes to serve. The market as a whole should take priority over the intended audience as much as possible. For example, the national market is half men, and although most men don’t buy women’s clothing, they should still be included so we can see the percentage of men that are potential buyers. Even if only 2% of men purchase women’s clothing, the effect on the TAM can be large.
Quotas for age, gender, income, race, ethnicity and geography are all useful for a TAM because the goal is to extrapolate the proportion of those who are potential customers to the greater market. These quotas will come from census data. We should typically use census data so we know our estimates are as accurate as possible. While it would be nice to interlock as many demographic variables as possible, the cost of the surveys may become exorbitant. If cost becomes a barrier, we can remove certain demographics that we believe may not significantly differ across the market. The interlocking quotas will serve as our groups for the TAM calculation.
Example: The TAM calculation will consist of dozens of groups that are created from the interlocking demographic quotas you create for the market (e.g., 18-24-year-old females making under 50k who live in the midwest). For each group, we need a large enough sample to be confident in extrapolating the sample percentage of buyers and annual spending on the product or service to the greater market.
Keep in mind that the sample for the markets they currently serve should also be sufficiently large to increase confidence in those specific estimates. It would be wise to consider oversampling in these markets.
As will be discussed, a willingness-to-pay module should typically be included for more accurate SAM estimates. Careful consideration of the sample size is needed for the WTP questions to be useful. Specifically, we need to have enough respondents in each quota group to run the necessary subgroup analyses to extrapolate the findings to the larger population. Beyond quotas, we should also weight the results back to the census data to ensure our sample is as representative of the larger market as possible.
For simplicity, this documentation will separate the survey questions needed to run a TAM and a SAM. In reality, many of the questions will be put in the same module as the questions needed for the SAM calculation are mostly follow-up questions from the TAM questions.
The first set of questions should contain demographic questions so we can accurately place respondents in our market quotas. As stated above, these questions will typically ask about age, gender, income, geographical location, race, and ethnicity.
Example: For a women’s clothing TAM on the national level, our quotas would largely look the same as they do for Trendlines. Therefore, the first set of questions will be our standard demographics.
The second set of questions should directly inform the TAM by asking questions that will help us categorize a respondent as a potential buyer or not. Unlike with our traditional surveys, we will want to keep terminated respondents, either in Qualtrics or through an API, so we can calculate the portion of potential buyers and non-buyers by quota. For the TAM, the focus should be on the general market, not on the specifics of the product.
Example: For our client who sells high-end dresses, our first goal is to see who purchases women’s clothing.
The third set of questions should address the frequency of purchases and money spent per purchase. Typically, we need enough information to extrapolate the purchase behaviors to annual expenditure. Again, our goal is to get responses for the general category of product at this stage.
Example: For our client who sells women’s clothing, we should ask respondents how often they shop for women’s clothing, and how much they typically spend during each shopping occasion. Several questions will typically be needed for triangulation. For the TAM, the questions should be focused on the purchasing habits for the broader category, as questions about the specific product will be included in the SAM (i.e., women’s clothing, not high-end dresses).
To gather information for the SAM calculation, we want to ask many of the same questions we did in the second and third sets of TAM questions, but with a focus on the particulars of our client’s product or services. Remember, the TAM sizes the market for the entire product category, and the SAM focuses on the specific product the client offers or will offer.
The first set of SAM questions will be used to determine whether the respondent is a potential buyer of the specific product or service our client sells.
Example: For respondents that purchase women’s clothing, we will then ask them whether they purchase dresses. Moreover, we will want to see whether a respondent purchases dresses in the price range and style that our client offers. Careful consideration is needed for the SAM questions. While someone may spend $1000 a year on dresses, none of that may be relevant if the respondent is not willing to pay the price of our client’s dress ($200) at one time. In other words, the questions for SAM need to match the actual purchase-making process offered by our client.
The second set of questions should address purchase frequency and money spent per purchase for items that are more specific to what the client sells. Instead of asking about the frequency of women’s clothing purchases, we now want to know about their habits for high-end dresses. There are many possible questions to consider, but a common one may be the purchase method. If our client only sells their clothing online, then we need to make sure the respondents are willing to purchase high-end dresses online.
The third set of questions is optional and will depend on the client’s research questions and budget. Whenever possible, we should include a willingness-to-pay (WTP) module for the client’s specific product or service. Currently, we are most prepared to assess WTP via a conjoint. When paired with a dual-response none, a conjoint can provide specific estimates for who is willing to purchase the client’s product or service, and at what price.