Leadership1 distinct publisher3 min readUpdated
J's Breakfast Club in Gary, Indiana traded at a tenth of normal for eight and a half days after a derecho. Its recovery kit was bought at retail, on day two.
The Board Room · Leadership desk

Compiled by The Board RoomSomething wrong?How this is made
J's Breakfast Club in Gary, Indiana sold roughly a tenth of its normal volume during the eight and a half days it spent without power after the August 11 derecho, a shortfall owner Joslyn Kelly puts at about $19,000 [1][2][4][5]. The storm itself ran for an afternoon; the outage ran for more than a week, and as of Thursday evening more than 57,000 Northern Indiana Public Service Co. customers were still without power [8]. The gap between the weather event and the service event is the part most operators have never costed.
The most expensive detail in Kelly's account is not the wind. She was told power would be back on Friday; Friday became Tuesday, and Tuesday became the following Tuesday [7]. Every revision is a staffing call, a food order and a reopening notice that has to be unwound, and it is why "wait for restoration" is not a plan.
Recovery began on the second full day, when Kelly took inventory and found the kitchen's gas still worked [9]. She bought tents, portable lights and propane-fuelled deep fryers and griddles from Menards, set up outside the building, and ran on power banks and candles [10]. Several employees turned up; corporate customers called with orders for staff still on site; she served food alongside the City of Gary at the YMCA [11]. That is a competent response. It is also one assembled after the fact and bought at retail under duress, and it still recovered only a tenth of normal trade [10][4].
The arithmetic is worth writing down. If $19,000 is about 90 percent of expected sales, normal takings for the period were near $21,100, or roughly $2,480 a day, with about $2,235 a day forgone [6]. That is the number a generator, a pre-negotiated propane supply or a standing arrangement with a powered kitchen has to beat on a per-day basis. Meanwhile the fixed side does not pause: Kelly notes she still had staff to pay, bills due and commitments to meet while unable to serve anyone [12].
Restoration is also not recovery. The restaurant reopened for dine-in and carry-out the day after power returned, but has not switched delivery back on, a channel that normally accounts for 10 to 15 percent of the business [13]. On that basis it is running at 85 to 90 percent of its usual channel mix with the lights on [14]. Kelly expects to keep missing revenue while customers around her remain without electricity [17]. Two small details capture the eight days: someone took the tent from her patio, and she learned power had returned when the alarm company called about motion inside the building [15].
What to watch is whether the utility's remaining restoration timelines hold, given that the ones given to Kelly moved three times [7][8]. Her own diagnosis is that Gary's infrastructure is very old [16]. If that is the operative variable rather than the derecho, then an eight-day outage is a planning baseline rather than an outlier, and the relevant question for any single-site operator is what a bridge across it costs per day against roughly $2,200 of daily sales [6].
Follow any of these and your For You feed starts watching them — no settings page required.
Ranked by verification strength, evidence, and original report placement.
Kelly says the restaurant sold about 10% of what it typically does during the blackout, losing about 90% of revenue, to the tune of about $19,000.
Kelly says Gary has very old infrastructure.
Joslyn Kelly is the owner of J's Breakfast Club in Gary, Indiana.
The restaurant lost power around noon on August 11 as a derecho, a widespread long-lived wind storm, passed through.
Power did not return until around 1 a.m. on Thursday, August 20.
If $19,000 represents about 90% of expected sales, normal takings for the outage period were about $21,100, or roughly $2,480 per day, with about $2,235 per day of sales forgone.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Single-source first-person account
Every substantive claim traces to one as-told-to essay from one publisher, edited for length and clarity. The narrative is internally consistent and specific about dates, equipment and channel mix, and it carries one third-party-attributed datapoint (the NIPSCO customer count). But the central financial figure is self-reported with no records, the utility is not given voice on restoration, and no other Gary business is compared, so the evidence base cannot be triangulated.
One documented site, real but unreplicated
There is concrete, dated evidence of actual use rather than intention: a retail off-grid kit deployed on day two, employees working, paying corporate orders, a city/YMCA community feed, and a staged reopening with delivery still off. That is genuine operational adoption, but it is a single site at roughly 10% of normal throughput, with no evidence that other businesses in the 57,000-plus affected footprint did the same or that the practice persists post-restoration.
Modestly overstated by generalization
The account itself is restrained and specific — approximate figures, acknowledged uncertainty, no product or vendor being sold. The overstatement sits in the framing around it: a single self-reported restaurant loss is presented as evidence that 'the small-business risk is now the grid, not the storm', a structural claim that one first-person anecdote and one customer-count datapoint cannot carry. The gap is generalization, not exaggeration of the underlying facts.
Sympathetic first-person format, no disclosed commercial stake
The subject is the affected business owner and the sole narrator, with clear interest in a sympathetic account, in pressure on the utility and local government, and in visibility for a reopening restaurant; the as-told-to format means the publisher relays rather than tests those claims and the utility gets no reply. Against that, no product, vendor, funding round or paid placement is being promoted, the losses are framed as approximate, and the operational detail is mundane rather than self-flattering — so the incentive to distort is moderate, not severe.
Plausible and internally consistent, thinly sourced
Confidence is limited by structure rather than by contradiction: nothing in the cluster conflicts, the arithmetic on the reported loss checks out against the stated outage window, and the operational detail is the kind that is hard to fabricate. But with one publisher, one narrator, no records and no utility response, the specific magnitudes should be treated as indicative and the wider structural conclusions as unproven.
Distinct publishers with included, body-backed reporting in this cluster.