Discovery

Find the work worth changing.

We sit with the people who drive revenue and the people who carry cost. Not a top ten executive list. Not a survey. Department by department, workflow by workflow, until we can see the business the way the people inside it see it. Then we rank what we found, and we start building.

01 1 conversation

The founder interview comes first

We start with the person who owns the outcome. One conversation, founder or CXO, no agenda but the business itself. Where revenue actually comes from. What the cost structure looks like, line by line. Where the founder's own time goes every week. That conversation sets the map. Every interview after it gets checked against what we heard here.

02 15–20 interviews · ~2 weeks

Then every department, over about two weeks

15–20 interviews, one at a time Day 1 Day 14

From there we go wide. One knowledgeable person per department, marketing, B2B, post-sale, CRM, tech, each sitting at a different distance from revenue. Fifteen to twenty interviews, about two weeks. Nobody fills out a form. Everybody talks to us directly, about their own work, in their own words.

Marketing B2B sales Post-sale and support CRM Tech
03 One level below the job title

We map the workflow, not the org chart

For every department, we go one level below the job title. What tools this person actually uses. Which steps are manual, which are already automated. How much time each workflow costs, every week. How close that workflow sits to revenue or cost. That last question matters most. A slow workflow next to revenue is worth more to fix than a slow workflow three steps removed from it.

Job title ↓ one level down
01 Tools in use, not tools on the license list
02 Manual steps versus automated steps
03 Time cost per workflow, per week
04 Distance from revenue or cost
04 Few wedges, ranked

Rank on time and revenue, not on who complained loudest

3–5 wedges, ranked

Every workflow we find gets placed on two axes: how much time it costs, and how close it sits to revenue or cost. The problems that score high on both get built first. This is the point of discovery. Not a long list of everything that is broken. A short, ranked list of what to fix first, and the reasoning for why.

Discovery map, illustrative
Revenue proximity
Fix first Low High High Low Marketing B2B sales Post-sale Marketing CRM CRM Tech Tech
Time cost, today
High time cost, high revenue proximity: built first
Everything else: queued, not ignored

Illustrative. Every engagement produces its own map, from its own interviews.

05 Runs underneath everything else

A bounty board and a hackathon, running underneath everything else

Interviews are a two week snapshot. Problems keep showing up after we leave. So we set up two things that keep surfacing them. One runs every day. The other runs a few times a year and forces out what the day to day misses.

05

The bounty board

Keep surfacing problems after the interviews end · runs every day

A manual step, repeated weeklySubmitted
Two tools, no sync between themClaimed
A report built by hand, every FridayApproved

Employees submit daily problems as bounties. Others pick them up and solve them. A manager approves. A small perk rewards it. The result is a live, validated database of real problems, built by the people doing the work.

06

The hackathon

Force out problems the interviews missed · a few times a year

Outcome-based hackathons surface issues nobody thought to mention, in ops, in merchandising, in supply chain, wherever the team points it. It is a fast way to test whether the discovery map is actually complete.

The difference

Not a report. The start of a build.

A consultant

Interviews ten to twenty executives. Runs a survey for everyone else. Hands over a deck, in about the same two weeks this whole discovery phase takes.

Celeco

  • Interviews reach the people doing the work, not just the executives above them
  • Discovery ends in a working system, not a slide deck
  • We embed in existing workflows instead of introducing new tools
  • We show a live demo on the business's own problem before the build starts
  • We stay through deployment and governance

There is no shelf for recommendations to sit on, because there is no report. There is a system that runs.

Fifteen to twenty conversations. About two weeks. One ranked list at the end.