Lubinpla

Measuring Technical Service Desk Time Spent on Routine Questions

Summary

Measuring how much technical service time goes to routine questions takes four steps: define routine, separate document requests from judgment questions, count from records, and rank the fixes. No published benchmark gives that share for chemical technical service. The figure only means something for your own team, over a fixed period, with the definition written down before anyone counts.

  • EU REACH requires a safety data sheet to be provided free of charge, on paper or electronically, no later than the date the substance or mixture is first supplied (REACH Article 31(8), consolidated 2026).

  • US OSHA requires wholesale distributors to provide safety data sheets to employers or other distributors upon request (OSHA 29 CFR 1910.1200(g)(7)(vi), 2024).

  • ITIL practice, as summarized by Atlassian, treats a service request as a formal request for something to be provided, distinct from an incident, which is an unplanned disruption to a service (Atlassian, 2026).

  • One help desk vendor, eesel, defines a repeat contact rate as solved tickets where the same customer came back about the same issue within 7 days, divided by all solved tickets (eesel, 2026).

What counts as a routine question in technical service?

A routine question in technical service is one that a written answer, an existing document or a fixed rule already settles. The engineer's work is finding the answer and sending it, not deciding anything. Write that definition down, with examples, before anyone counts; otherwise each engineer draws the line in a different place.

For a chemical supplier, routine usually covers three groups (illustrative): document requests, such as a safety data sheet (SDS), technical data sheet (TDS), certificate of analysis (CoA) or regulatory certificate; status questions, such as where a sample or an order stands; and repeat answers, where another customer has already received a written answer to the same product question.

IT service management draws a similar line. ITIL practice, as summarized by Atlassian, treats a service request as "a formal request from a user for something to be provided, as distinct from an incident" (Atlassian, 2026), and notes that "most service requests recur" (Atlassian, 2026). A technical service team can borrow the distinction, but not the IT categories. A password reset has no counterpart in chemical supply, and a request for a CoA for one lot has no counterpart in IT.

If the answer depends on the customer's conditions, such as substrate, temperature, contamination or process history, the question stays non-routine. Complaints stay out of the routine count as well, because ISO 10002:2018 treats complaints handling as its own process, including analysing and evaluating complaints to improve products and services (ISO, 2018).

How do you separate document requests from questions that need technical judgment?

To separate document requests from technical judgment questions, ask one thing of each case: could someone with access to the document library have closed it without deciding anything? If yes, it is a document request. If the answer required a reading of the customer's conditions, it is a judgment question, even when a document went out with the reply.

Many cases mix the two. A customer who asks for the TDS "and whether it will work at 150 degrees Celsius" (illustrative) has made one document request and one judgment question. Code the case by its hardest part, which here is judgment, and record that it also contained a document request so the count of document requests stays complete.

Some document requests exist because a rule requires the document. EU REACH requires an SDS to be provided "no later than the date on which the substance or mixture is first supplied" (REACH Article 31(8), consolidated 2026). In the US, a chemical manufacturer or importer must also provide distributors or employers with an SDS upon request (OSHA 29 CFR 1910.1200(g)(6)(iv), 2024). So a high count of SDS requests from customers who already buy the product suggests the SDS is not reaching them with the shipment. Cut those requests by fixing delivery, never by slowing or refusing them.

How do you count technical service time on routine questions from existing records?

Count from the records you already keep, the case log, email and the help desk, and not from memory. The work takes five steps.

  1. Fix the period and the list of channels, such as two full months of email, phone notes and help desk tickets (illustrative), then freeze the case list.

  2. Merge contacts into cases. One customer problem is one case, however many emails, calls or tickets it produced.

  3. Code each case as document request, status question, repeat answer or judgment question, using the written definition.

  4. Take handling time from the records. Use the logged time where the help desk records it; otherwise multiply the count of replies and calls per case by a fixed allowance per touch, and label the result an estimate.

  5. Divide routine handling time by all handling time, and state the denominator and the cases left out.

Merging matters because help desks split cases. In Zendesk, a reply to a solved ticket after the automatic close of 4 days becomes a new follow-up ticket (eesel, 2026), so one document request can show up as two tickets. The same discipline on stages and denominators appears in how to find where technical support cases get stuck, which applies it to waiting time rather than question type.

How do you check a routine-question time measurement?

A routine-question time measurement has to pass six checks: a written definition, a split between document and judgment questions, time from records, merged duplicates, a stated denominator and ranked reductions. A result that fails any one of them cannot be compared with the next measurement, so nobody can tell whether a fix worked.

ISO 9001:2026 requires organizations to monitor, measure, analyze and evaluate the performance and effectiveness of their quality management system (ISO, 2026), and a measurement that cannot be repeated does not serve that purpose. Run the check before sharing the result, because a share that circulates once gets quoted for years.

Use the table below to decide whether a routine-question time measurement is ready, and what to fix when it is not.

Check

Pass

Fail and what to do

Routine definition

Written with examples before counting

Defined after reading the cases: write it, then recode

Document requests split

SDS, TDS, CoA and certificate requests coded apart from judgment questions

Mixed cases counted as routine: code by the hardest part

Time source

Logged time or reply counts from records, method stated

Engineers' estimates: replace with record counts

Duplicates

Follow-ups and cross-channel contacts merged into one case

Ticket count used as case count: merge linked tickets

Denominator

All handling time in the period, exclusions listed

Share of an unknown total: state the base and exclusions

Reductions

Ranked by hours saved, each with an owner

A list without owners: assign one per fix

Of the six rows, the time source decides whether the share is real. An estimate from memory tends to reflect the most irritating questions rather than the most frequent ones, so take time from records or say plainly that the figure is an estimate.

Which routine technical questions should you reduce first?

Start with the routine questions that take the most hours in total, counting both how often they arrive and how long each takes, and that one owner can remove with one change. Rank by hours saved, not by how often engineers complain about a question.

Document requests are usually the easiest to remove, because the fix sits upstream, before the customer has to ask. If customers ask for the SDS after delivery, check that it went out with the initial shipment and with the first shipment after each update, as US OSHA requires of chemical manufacturers and importers (OSHA 29 CFR 1910.1200(g)(6)(i), 2024). A CoA requested for every lot belongs in the delivery documents or the shipping notice. A TDS request means the current version should be published where customers can find it.

Status questions shrink when customers get the next update before they ask. The eesel analysis of repeat tickets lists "no update while waiting" among the four causes that "cover most" repeat contacts (eesel, 2026).

Repeat answers shrink when the written answer becomes a short note or an FAQ entry that every engineer sends the same way.

Lubinpla's technical support performance analysis counts inquiry records against definitions fixed in advance, then diagnoses where the bottlenecks are and which improvements come first, in one document for review. To see whether each fix took hours off the desk, recheck the share over a period of the same length, with the same definition.

Common mistakes when measuring routine technical service questions

The mistakes that come up most are estimating instead of counting, counting tickets instead of cases, and treating every document request as wasted time.

  • Estimating time from memory. Engineers remember the irritating questions, not the frequent ones. Count from records, even roughly.

  • Counting tickets instead of cases. Follow-up tickets, and the same request sent by email and by phone, inflate the routine share. Merge first.

  • Calling a mixed case routine. A TDS request with a performance question attached is a judgment question. Code it by the hardest part.

  • Dropping required documents. An SDS has to reach the customer, so the fix is earlier delivery, not refusing or delaying the request.

  • Changing the definition between measurements. A new rule for what counts as routine moves the share without any real change. Keep the text fixed and date any change.

  • Quoting an outside percentage. No published benchmark for chemical technical service supports one, so the only meaningful comparison is your own team over time.

Frequently asked questions

Is a high routine share a sign of a weak technical service team?

No. A high routine share usually means documents, status updates or standard answers are not reaching customers on their own. It points to gaps in delivery and publishing, which other teams often own, more than to the engineers' skill.

Should routine questions move away from technical service engineers?

Move them only after the upstream fix has been tried. Routing document requests to customer service still costs time, while sending the documents with the shipment removes the request. Keep judgment questions with engineers, and give whoever takes the routine questions the written answers to use.

How long a period should a routine-question measurement cover?

Cover at least one full business cycle that includes your normal mix of orders and samples, and use the same length each time you measure. Seasonal products need one period in season and one out of season; otherwise the routine share will reflect the season.

Related posts