How you will know whether it worked.
Agreed before launch, measured from a recorded baseline, reported every week against numbers you chose. This page sets out the mechanism — and points at the documents that show it running.
Four steps that make a result provable.
Most outsourcing engagements cannot prove they worked, because nobody recorded what "before" looked like. This is the sequence that fixes that.
-
Step 1
Record the baseline before anything moves
Your current response times, resolution times, accuracy, throughput and satisfaction — captured from your systems during Discover, while your team is still doing the work.
- If the data is unreliable, we say so rather than using it
- The baseline is agreed in writing by both sides
- It is never reconstructed afterwards from memory
-
Step 2
Agree the measures, not just the targets
Two organisations can calculate "resolution time" three different ways. The definitions are written down before the number matters.
- We report in your definitions, not ours
- Every metric names its source system
- What is excluded is stated as explicitly as what is included
-
Step 3
Calibrate before committing to a target
A target set before anyone has run the work is a guess. We commit to activity and quality standards immediately, and to outcome targets after a calibration period.
- Typically four to six weeks past the end of ramp
- The target is then contractual
- A target we think is unreachable gets said out loud, early
-
Step 4
Report weekly, including the bad weeks
A fixed day, a fixed format, every metric against baseline and target — plus what missed, why, who owns the fix and when it gets re-checked.
- You get access to the underlying data, not just the summary
- Misses appear in the pack, not in a conversation you have to start
- The actions log carries everything forward until it is closed
Every number against a baseline and a target.
A sample week from a support account. The comparison columns are the point: a number on its own tells you nothing about whether the engagement is working.
See the method, not a description of it.
These four are the closest thing to evidence we can give you before there is a published client engagement — and in some ways they tell you more.
Sample reporting pack
A full week including two SLA misses, their causes and their owners.
Read itQuality scorecard
Six weighted criteria, two of them contact-zeroing, with their consequences.
Read itA worked SOP
Nine steps, six exceptions, and a version history that shows it being maintained.
Read itTransition plan
Five tranches, each with an entry condition and a rollback trigger.
Read itA number without a baseline, a period and a source is just a bigger font.
When Novatek publishes engagement results they will carry all three, with the client’s written approval. Until then, ask us for a reference conversation — it is worth more than a case study anyway.
Ask for a conversation instead.
A written case study is a curated artefact. A twenty-minute call with someone running a comparable operation is not, and it answers the questions you actually have.
Ask us early in the evaluation. How quickly and how specifically a provider answers a reference request tells you a great deal — including when the answer is that they cannot arrange one yet.
What to ask a reference
- What went wrong in the first ninety days, and how was it handled?
- How long did it really take to reach the standard you wanted?
- What do you have to spend time on that you expected not to?
- Has the reporting ever told you something you did not want to hear?
- What would you do differently if you started again?
- Would you expand the scope, and if not, why not?
These are better questions than "are you happy with them", and we would rather you asked them.
Bring us a number you want moved.
Response time, accuracy, throughput, cost per contact, meetings booked — tell us which one matters and we will show you how it would be baselined, targeted and reported.