Black Friday Dashboard: What the Business Owner Should Watch

Key takeaways

• The unanswered queue is the first number of the day: according to the Harvard Business Review (2011), replying within 1 hour gives about 7 times the chance of qualifying the lead.
• Count as a sale only the payment confirmed at the source; proposal sent and "about to close" go in another column.
• ROAS compares attributed revenue with ad spend; it does not deduct product, discount, fees, or returns, so it is not profit.
• Always compare the same period with the same population: Friday against Friday, not against the whole week.
• When the sale's source is not available, the dashboard shows "no source"; inventing a source destroys the per-ad reading.

Why does the dashboard need to separate demand, capacity, sales, and result?

Because each calls for a different decision. High demand with low capacity calls for people; high sales with a low result call for a discount review; high clicks with low demand call for an ad review. A dashboard that lumps everything into "leads" hides which decision to make.

According to Confi Neotrust (2025), Black Friday brought in R$ 10.19 billion in Brazilian e-commerce, with 21.5 million orders and a falling average order value. More and smaller orders mean more conversations per real (R$) sold; a dashboard that does not show queue and order value does not show what is happening.

Decision rule: if a number on the dashboard does not lead to an action by someone with a name, take it off the main screen. A Black Friday dashboard is for deciding, not admiring.

The right-now view: the three alerts that demand immediate action

The right-now view has three alerts, each with an owner and a deadline: unanswered queue, stalled negotiations, and orders with exceptions.

1. Unanswered queue: conversations started by the customer with no human or AI reply. Alert when any conversation passes 15 minutes. HBR (2011) found that 24% of companies took more than 24 hours to reply and 23% never replied; on Black Friday, those conversations go to the competitor.
2. Stalled negotiations: conversations with a proposal or link sent and no reply for more than 6 hours. Alert for the owner to re-engage with the recorded pending issue.
3. Orders with exceptions: payment confirmed with no stock, incomplete address, link error. Alert for whoever runs operations, not the salesperson.

If the unanswered queue grows for two straight hours, the decision is about capacity, not messaging: redistribute the team or switch on the AI's automatic mode. The calculation of how many people are needed is in how to size your customer service team.

The period view: the eight numbers that summarize the campaign

The period view shows, per day and cumulative, eight indicators with a fixed formula. The ranges below are practical references, not market benchmarks.

A CRM with attribution, such as RASTRO, calculates revenue and ROAS per campaign, ad set, and ad from the sale recorded in the conversation; it is up to the team to record the value and confirm the payment, as described in from payment link sent to payment confirmed.

Why is ROAS not profit, and how do you calculate the acceptable minimum?

ROAS is the revenue attributed to the ad divided by the ad spend. A ROAS of 5 means R$ 5 in revenue for every R$ 1 invested; it does not mean R$ 4 in profit, because the revenue still pays for product, fees, shipping, and returns. The number the owner needs is the minimum ROAS, calculated from the contribution margin:

• Contribution margin per sale = discounted price − product cost − payment fees − shipping paid by the company.
• Contribution margin (%) = margin per sale ÷ discounted price.
• Minimum ROAS = 1 ÷ contribution margin (%).

Decision rule: if the ad's ROAS is below the minimum, each additional sale increases the loss; pause or reduce. If it is above, the ad can scale as long as there is stock and reply capacity. How to set the discount without crushing margin is in discounting without destroying your margin; reading the ad beyond ROAS comes down to knowing whether it is actually turning a profit.

Hypothetical example: conversations doubled and sales barely rose

The numbers below are hypothetical and serve to show the calculation, not an expected result.

1. Thursday. 100 conversations, 10 confirmed sales, average order value R$ 200, ad spend R$ 500. Conversion = 10 ÷ 100 = 10%. Revenue = R$ 2,000. ROAS = 2,000 ÷ 500 = 4.0.
2. Friday. 200 conversations, 12 sales, order value R$ 180, ad spend R$ 1,200. Conversion = 12 ÷ 200 = 6%. Revenue = R$ 2,160. ROAS = 2,160 ÷ 1,200 = 1.8.
3. Margin. Product cost R$ 90, payment fee 4% (R$ 7.20 on R$ 180), shipping paid by the company R$ 15. Margin per sale = 180 − 90 − 7.20 − 15 = R$ 67.80. Margin (%) = 67.80 ÷ 180 = 37.7%. Minimum ROAS = 1 ÷ 0.377 = 2.65.
4. Reading. On Friday the ROAS of 1.8 was below the minimum of 2.65: each sale lost money after the ad. Revenue rose 8%, but conversion fell from 10% to 6%.
5. Decision. Before raising the budget, investigate the queue: if reply time exceeded 15 minutes, the problem is capacity; if it did not, it is the audience or the offer.

How to compare numbers without fooling yourself

Compare the same interval, the same population, and the same sale criterion; log every change in price, budget, or campaign with date and time.

• Same interval: Friday against the previous year's Friday, or against the average of the four previous Fridays. According to Confi Neotrust (2025), before Black Week e-commerce was already growing about 33% because of early offers; comparing with the previous week distorts the effect.
• Same population: conversations that came from ads against conversations from ads; mixing in old customers changes conversion without changing anything in the ad.
• Same criterion: sale confirmed at the source on every day. If on Thursday the team counted "about to close" and on Friday only payments, the drop is artificial.
• Change log: a simple log (date, time, what changed) explains later why conversion fell at 2 p.m.

Common mistakes on the Black Friday dashboard

• Counting a click or a conversation as a result. A click measures interest in the ad; a conversation measures interest in the offer; only payment measures a sale.
• Presenting ROAS as profit. The business partner decides to raise the budget based on a number that deducts no cost at all.
• Attributing a sale with no source to an ad by guesswork. The per-campaign reading stops being reliable and the budget goes to the wrong ad.
• Looking at the day's average instead of the queue right now. An average reply time of 10 minutes hides the hour when 40 people waited 50 minutes.
• Having ten charts and no owner. A number with no owner generates no action.

When should the business owner look at the dashboard?

At three fixed moments per day and whenever an alert fires. Looking every five minutes leads to decisions driven by noise, such as pausing an ad over one weak hour.

• Morning (before customer service opens): queue accumulated overnight, orders with exceptions, stock of the items on offer.
• Midday: reply time, proposals sent, sales confirmed so far, ROAS per ad to decide on budget adjustments.
• End of day: confirmed revenue, conversion, sales with no source, the day's change log, and the decision for the next day.

Alerts go to whoever can act: the queue to the customer service coordinator, order exceptions to operations, ROAS below the minimum to whoever manages the ad. These moments go into the 30-day plan.

Frequently asked questions

Sources

• Confi Neotrust, Black Friday 2025 results
• Harvard Business Review, The Short Life of Online Sales Leads (2011)