The Department of Public Utilities (department) has allowed the Bay State Gas Company (company) to file new rates to produce more than $8,000,000 in additional annual revenue. The company proposed to allocate the increase among residential, commercial and industrial customers by a “rate base distributor” method based on a cost of service study prepared by Gilbert Associates, an independent consultant. Monsanto Company (Monsanto), an intervener, proposed an alternative allocation based on the same study. The department rejected both allocation proposals, and ordered that the rates for all customer classes be increased by a proportional percentage. Monsanto appealed, and a single justice of this court reserved and reported the case to the full court. The sole issue argued is the allocation of the increase. We affirm the decision of the department.
The Gilbert study allocated revenues and costs of service to the three main classes of customers, and calculated net operating income for each class. See Jones, Judicial Determination of Public Utility Rates: A Critique, 54 B.U.L. Rev. 873, 885-890 (1974). The rate base was also apportioned among the three classes, and a rate of return on rate base was calculated for each:
Residential 8.1%
Commercial 13.8
Industrial 19.7
Total Company 10.0
A company witness testified that the rate of return for each class should be “within an acceptable band around the Company’s overall rate of return.” In the present case, the *319company and Monsanto argue, residential customers have been “subsidized” by commercial and industrial rates of return.
To bring the rates of return for the three classes closer together, the company proposed to allocate the rate increase in proportion to the allocated percentages of the company’s total rate base. Monsanto proposed more drastic equalization. The two proposals would have produced the following allocations of the rate increase:
Company Monsanto Residential 76.6% 87 %
Commercial 17.3 13
Industrial 6.1 None
In support of its proposal, the company also submitted data to show that there was a danger that some industrial customers might convert from gas to oil, but that residential gas rates were lower than residential rates for oil or electricity.
The department stated three grounds for rejecting the company proposal. First, the Gilbert study calculated rates of return for 1976, and did not take into account subsequent changes, while the department’s decision was based on a 1977 test year. Second, allocation of the rate increase solely on the basis of rate base, excluding operating expenses, would be inconsistent with the department’s determination of required revenue, which involved adding operating expenses to the allowed return on rate base. Third, the department was not convinced that the company was in danger of losing industrial customers.
No claim of confiscation is before us, and the standard of review is governed by G. L. c. 30A, § 14 (7). Boston Edison Co. v. Department of Pub. Utils., 375 Mass. 1, 9 (1978). Within a substantial range, business decisions such as allocation of rates among specific services are matters for the company’s determination, but even in such matters the company when challenged must come forward with evidence to explain its decisions. Fitchburg Gas & Elec. Light *320Co. v. Department of Pub. Utils., 375 Mass. 571, 578-579 (1978). New England Tel. & Tel. Co. v. Department of Pub. Utils., 371 Mass. 67, 83-84 (1976). As Monsanto recognizes, although an intervener’s case is not to be ignored, it is not the focus of attention. Trustees of Clark Univ. v. Department of Pub. Utils., 372 Mass. 331, 334-335 (1977).
There is here no showing that the department has departed from a rule previously followed, as in New England Tel. & Tel. Co. v. Department of Pub. Utils., 371 Mass. 67, 84 (1976), and Boston Gas Co. v. Department of Pub. Utils., 367 Mass. 92, 103-105 (1975). Rather, the company seeks to reallocate revenue burdens among classes of customers, and the burden is on the utility to show that the reallocation is proper. See Metropolitan Dist. Comm’n v. Department of Pub. Utils., 352 Mass. 18, 24 (1967); Fitchburg Gas & Elec. Light Co. v. Department of Pub. Utils., 371 Mass. 881, 885 (1977). Allocation on the basis of cost, including both operating expenses and return on investment, is well recognized. See Jones, Judicial Determination of Public Utility Rates: A Critique, 54 B.U.L. Rev. 873, 884-885 (1974); J. Bonbright, Principles of Public Utility Rates 294 (1961); New England Tel. & Tel. Co. v. Department of Pub. Utils., 371 Mass. 67, 82-85 (1976). But cost need not be the sole criterion. Boston Edison Co. v. Department of Pub. Utils., 375 Mass. 1, 46 n.29 (1978). Cf. G. L. c. 164, § 94G (fuel charges to be billed at single uniform rate). “When alternative methods are available, the Department is free to select or reject a particular method as long as its choice does not have a confiscatory effect or is not otherwise illegal.” Massachusetts Elec. Co. v. Department of Pub. Utils., 376 Mass. 294, 302 (1978).
Applying these principles, we conclude that there was no error. Doubtless it was open to the department to decide that residential customers had borne less than their fair share of the company’s revenue needs in 1976, and that there was no indication of any great change in this respect in 1977. See Jones, Judicial Determination of Public Utility Rates: A Critique, 54 B.U.L. Rev. 873, 885 (1974). Simi*321larly, the department might have accepted the company’s evidence that it was in danger of losing industrial customers. But we cannot say that such decisions were compelled. As to changes from 1976 to 1977, the company witness gave only general conclusions. As to loss of customers, there was evidence that from 1976 to 1977 the company lost 3.9% of its residential sales and only 2.8% of its commercial and industrial sales.
More significant is the company’s failure to take account of operating expenses in its proposed allocation. Using the Gilbert study, the department calculated that allocation of the rate increase in accordance with the total of operating expenses plus return on rate base would produce the following:
Residential 67.1%
Commercial 20.8
Industrial 8.6
Special contracts 3.5
The company was required to prove its case before the department by presenting a clear and reasonable analysis. Fitchburg Gas & Elec. Light Co. v. Department of Pub. Utils., 375 Mass. 571, 582 (1978). We think the department could properly conclude that it was not required to omit operating expenses from the calculation. The proportional percentage allocation ordered by the department will produce the following:
Residential 64.0%
Commercial 22.3
Industrial 9.9
Special contracts 3.8
Comparing the two allocations, the department could reasonably conclude that the disparities were not such as to require it to devise some novel formula not presented by the parties.
*322The case is remanded to the Supreme Judicial Court for the county of Suffolk for entry of a judgment affirming the order of the department.
So ordered.