Formula to calculate the cost of a reopened conversation versus one resolved on first contact

What you're measuring (and what you're not)

A conversation resolved on first contact is one that ends with the customer without them needing to write again about the same issue. A reopened conversation is one the customer picks up again later, whether in the same WhatsApp thread or by opening a new one. What changes between the two is not the cost per message: it's the number of times the same case passes through your team. The cost per resolved conversation is covered on another page; here the focus is different: how much more expensive each extra contact becomes when the first one wasn't enough.

The usual trap is measuring messages. A reopened case can have few messages and still consume two or three full contacts, with two or three different agents, and with the customer angrier each round. The unit of measure that works is the contact: a service window with a beginning and an end. If the customer writes on Monday, leaves satisfied and comes back on Thursday about the same thing, that's two contacts, not one long one.

The four costs that go into the calculation

Before the formula, it's worth separating the components, because almost all calculations that go wrong do so because they mix costs that behave differently. There are four:

• Agent cost: the time the team spends on the contact, multiplied by the loaded cost per hour (salary plus social charges plus overhead, plus absenteeism and non-productive time). If you don't have that number, the formula to calculate an agent's cost per hour is on another page of the site.
• Context cost: the extra time it takes to understand a case that was already handled before. The new agent has to read the history, ask again what was already asked and reconstruct what was done. It's estimated as a percentage of the contact time, not as a fixed value.
• Dissatisfaction cost: the customer who comes back tends to write more, compare, ask for a supervisor or simply leave. It isn't measured directly in pesos, but it can be bounded by the drop in repurchase or campaign response for that segment.
• Opportunity cost: the agent time spent rehandling an old case is time that wasn't used to handle a new case. During peak hours this translates into queue, and the queue has its own formula.

The first two are the ones calculated precisely. The last two are estimated and reported separately, so as not to contaminate the hard number. If you put them all into a single figure, the figure stops being defensible and nobody uses it to decide.

The base formula, step by step

1. 1Define the loaded hourly cost of your team (HC). It is the same number you would use for any customer service cost calculation.
2. 2Measure the average time of a contact resolved on the first contact (T1), in minutes.
3. 3Measure the average time of a contact that reopens a previous case (T2), in minutes. T2 is usually greater than T1 because it includes re-reading history and revalidating data.
4. 4Calculate the reopening rate (R): out of every 100 closed cases, how many come back for the same issue. It is measured on cases, not on messages.
5. 5Cost of the case resolved on the first contact: C1 = (T1 / 60) × HC.
6. 6Cost of the case that is reopened once: C2 = (T1 / 60) × HC + (T2 / 60) × HC × (1 + k), where k is the extra percentage of context that the second contact consumes.
7. 7Difference per case: D = C2 − C1. That is the number you are going to multiply by volume.
8. 8Monthly impact: D × number of cases reopened per month.

The k factor is the most discussed and the one that changes the result the most. A k of 0.2 means that the second contact consumes 20% more time than a normal one because context has to be rebuilt. A k of 0.5 already implies that the agent spends half of the contact understanding what happened before. It is measured by watching three or four real cases with a timer, it is not estimated from memory.

Table to build the calculation with your data

This table is the minimum spreadsheet. Each row is a data point that you have to obtain or measure, with the unit and where it comes from. Do not fill it with estimates: if you do not have a data point, leave it blank and note how you are going to get it.

Numeric example with toy data

To see how the formula moves, let's use round numbers that are not meant to be realistic, just to show the mechanism. Let's assume CH = 4,000 pesos per hour, T1 = 12 minutes, T2 = 18 minutes, k = 0.3, and R = 22 cases per 100 closed. With those values:

• C1 = (12 / 60) × 4,000 = 800 pesos per case resolved on first contact.
• C2 = 800 + (18 / 60) × 4,000 × 1.3 = 800 + 1,560 = 2,360 pesos per reopened case.
• D = 2,360 − 800 = 1,560 pesos difference per case.
• If you have 400 closed cases per month and R = 22%, that's 88 reopened cases. Monthly impact = 88 × 1,560 = 137,280 pesos.

The number that matters is not the total: it's the relationship between D and C1. In this example, reopening costs almost three times what it costs to resolve on first contact. That relationship is what justifies or doesn't justify investing in improving the first response. If the relationship is 1.1, improving the first response barely moves the needle and it's better to look at another process.

How to measure the reopening rate without counting wrong

The reopening rate is the data point that gets unintentionally falsified the most. Three frequent mistakes:

• Count messages instead of cases. A customer who writes six times about the same thing is one reopened case, not six.
• Count reopenings within the same window as if they were a new contact. If the customer keeps writing ten minutes later, it is the same contact. You have to define the cutoff yourself and apply it consistently.
• Do not separate a reopening for the same topic from a new contact for a different reason. If the customer comes back for something else, it is a new case and does not count for this metric.

The cleanest way to measure it is to tag each closed case and check whether the same contact reappears about the same topic in the following days. The review window depends on your business: in an e-commerce it can be a week, in a service with appointments it can be a month. What matters is that the window is the same for all cases and is written down.

When first contact is not the right answer

Resolving on first contact is not always the cheapest option. There are cases where forcing a quick close guarantees a reopening, and there the cheap first contact turns out expensive. Three typical situations:

• Cases that need information the customer does not have on hand. Closing without that information guarantees they will come back.
• Cases that depend on a third party (a supplier, an internal area, an appointment). First contact cannot resolve them even if it tries.
• Complaint cases with an expectation of compensation. Closing quickly without resolving the underlying issue generates a much longer second contact with a higher risk of escalation.

In those three cases, the cost of a first contact that does not close is not C1: it is C1 plus the high probability of C2. The formula stays the same, but R spikes and the result changes. There the decision is not to rush the close, it is to route well and tell the customer when they will get an answer.

What to do with the number once you have it

1. 1Calculate D and the D / C1 ratio with your current data. If the ratio is less than 1.5, improving the first response is not the priority.
2. 2If the ratio is high, break down the reopenings by reason. The most frequent reason is where the concrete improvement lies.
3. 3Measure again after thirty days with the same formula. If the reopening rate dropped, the monthly impact dropped in the same proportion.
4. 4Leave the calculation written down with the date and the assumptions. A number without assumptions cannot be checked again and gets discarded.

The most common mistake is not calculating wrong: it is calculating once and not measuring again. The reopening rate changes with seasonality, with product changes and with team changes. A calculation from six months ago is no good for deciding today.

Limits of this formula

Three things this calculation does not capture and that are worth saying out loud. First, it does not measure the cost of dissatisfaction beyond the extra time: the customer who leaves and does not come back does not appear in the reopening rate, they appear in the drop in repurchase. Second, it does not distinguish by agent: if one agent has a much higher reopening rate than another, the average hides the problem. Third, it does not work for cases that resolve themselves, without human intervention: there the cost of the first contact is zero and the comparison does not apply.

With those caveats, the formula serves what it has to serve: putting a defensible number on the table when someone asks whether it is worth investing in improving the first response. Without a number, the discussion is a matter of opinion.