GB532 Unit 3 Secondary Data Review Example

Most of the Lost Gallons Were the Weather: A Unit 3 Secondary Data Review for a Propane Distributor

[Student Name]

Purdue University Global

GB532: Marketing Research

Unit 3 Secondary Data Review

[Instructor Name]

[Date]

The propane distributor, its records and every figure are composite and illustrative; the public data sources described and the sources cited are real.

Why Secondary Data First

The problem definition memo in Unit 2 asked how many customers heat partly with heat pumps and how they use propane. Stewart and Kamins (1993) argue that a researcher should exhaust existing data before collecting new data, because secondary sources are cheaper, faster and sometimes sufficient. They also warn that every secondary source was collected for someone else's purpose, so its fit with the current question must be judged. The source table below does that.

Exhibit 1: The Source Table

1. Internal delivery tickets. Collected by: Otter Creek's drivers and billing system. Why: to bill customers and schedule deliveries. Currency: every fill for ten years. Coverage: all automatic-delivery accounts, about 8,160. What it cannot tell: what equipment a home has, or why usage changed.

2. Heating degree-day records. Collected by: national weather records for the region's stations. Why: to measure heating demand. Currency: daily, current. Coverage: the distributor's service area. What it cannot tell: individual homes' behavior.

3. Federal propane price series. Collected by: the federal energy statistics agency's state heating oil and propane program, which reports residential prices during the heating season (U.S. Energy Information Administration, n.d.). Why: to track prices for consumers and policymakers. Currency: weekly in winter. Coverage: Vermont statewide. What it cannot tell: Otter Creek's own customers' price response.

4. Account start and stop records. Collected by: Otter Creek's customer service. Why: to manage accounts. Currency: current. Coverage: all accounts. What it cannot tell: reasons for leaving, which are recorded inconsistently.

5. Public summary of heat pump rebates by town. Collected by: the state's energy efficiency program. Why: to report program activity. Currency: annual. Coverage: towns, not households. What it cannot tell: which specific customers installed heat pumps.

The Key Measure: The K-Factor

Each delivery ticket records gallons and a date. Dividing the degree days between two fills by the gallons delivered gives an account's K-factor, the number of degree days a home gets from one gallon. Propane distributors use it to schedule automatic deliveries. A home whose K-factor rises sharply is using less propane for the same cold, which is what a heat pump would cause. The review computed every account's K-factor for the last two winters and flagged those whose propane use per degree day fell by at least a fifth: about 1,140 accounts.

Exhibit 2: Decomposing the Loss

Residential gallons fell from about 7.40 million to about 6.59 million, a loss of about 814,000 gallons.

Weather: degree days fell about 6.5 percent. About 90 percent of these homes' propane goes to heat, so the milder winter explains about 433,000 gallons, about 53 percent of the loss.

Accounts lost: about 94 net accounts closed, averaging about 790 gallons a year, about 74,000 gallons, about 9 percent.

Flagged accounts: the 1,140 accounts with sharply higher K-factors used about 190 gallons less each than weather alone would predict, about 217,000 gallons, about 27 percent.

Remainder: about 90,000 gallons, about 11 percent, spread across other accounts and consistent with modest conservation as prices rose during the winter.

More than half of the loss was the weather, which will change by itself; about a quarter points to heat pumps, which will not.

Do the Flagged Accounts Point to Heat Pumps?

The flagged group is a proxy, not a measure. Two checks support it. First, towns where the state's rebate summary shows more heat pump installations per thousand homes have more flagged accounts; across the 24 towns Otter Creek serves, the correlation is about 0.62. Second, the flagged accounts' reductions are largest in the shoulder months of autumn and spring and smallest in the coldest weeks, the pattern expected when heat pumps carry most of the load in moderate weather and propane takes over in deep cold.

Other explanations remain possible. Some flagged homes may have added insulation, replaced windows or had fewer people living in them. The secondary data cannot separate those cases from heat pumps.

Ten Years of Tickets

Ten years of delivery tickets also show the trend. The share of automatic-delivery accounts flagged for falling use was under 2 percent a year until four years ago, then rose to about 5, 8 and 14 percent in the last three winters. The rise matches the timing of larger state rebates for cold-climate heat pumps. If the trend continues, about a fifth of accounts could be flagged within two years. That projection is uncertain, since rebate budgets and equipment prices can change, but it shows why the decision about a backup-heat plan should not wait.

Quality of the Internal Data

Internal records have their own weaknesses. Drivers occasionally record a partial fill as complete, and a few accounts share a tank between two buildings. The review removed about 3 percent of tickets with impossible values, such as a fill larger than the tank's capacity, and checked that results did not change when those tickets were included. Account stop reasons were too inconsistent to use, which is one reason the survey must ask directly.

What the Prices Show

Residential propane prices in Vermont rose during the winter, according to the federal series. Higher prices may have encouraged conservation and made heat pumps more attractive. The price data cannot show how Otter Creek's customers responded, but they suggest that the remainder of the loss is plausibly price-driven.

What Secondary Data Cannot Answer

The review answers part of the problem: most of the loss is weather, and roughly a quarter is concentrated in accounts behaving as heat pump homes would. It cannot answer how many homes actually have heat pumps, how hybrid households use propane in the coldest weeks, what they would value in a backup-heat plan or whether they would buy installation services. Those are the questions that justify primary research. Malhotra (2019) describes the role of secondary data as narrowing and focusing primary research, and the K-factor flag does exactly that: it gives the survey a stratum of likely hybrid homes to sample.

Implications for the Next Steps

The research proposal should sample flagged and unflagged accounts separately, ask directly about heating equipment and when propane is used, and test interest in a backup-heat plan. The manager's original fear, that heat pumps are eating the business, is partly confirmed and partly overstated: the threat is real but about a quarter of last winter's loss, not all of it.

References

Malhotra, N. K. (2019). Marketing research: An applied orientation (7th ed.). Pearson.

Stewart, D. W., & Kamins, M. A. (1993). Secondary research: Information sources and methods (2nd ed.). Sage.

U.S. Energy Information Administration. (n.d.). State heating oil and propane program. U.S. Department of Energy.