The shape of a 90-day rollout
Field sales digitization projects rarely fail on technology. They fail on sequence: teams switch on dispatch consoles before beats are digitized, or hand reps an app before the item master is clean, and the field concludes — correctly — that the new system knows less than their notebook did.
What follows is a typical phased plan for taking a distribution or dealer field operation from paper beats to live dispatch over roughly ninety days. Treat the 90-day frame as a planning structure, not a guarantee — timelines vary with data quality, team size, and how much of the operation you bring into scope. The sequencing, however, holds across almost every rollout: master data, then routes, then reps, then dispatch. Each phase creates the ground the next one stands on.
Weeks 1–2: Master data — the unglamorous foundation
Everything downstream is a projection of three masters: outlets, items, and prices. Get these right before anyone sees an app.
- Outlets and dealers. Consolidate your outlet universe from whatever it lives in today — distributor lists, old spreadsheets, rep memory. Expect duplicates and dead outlets; plan the cleanup as part of the load, not after it. In xMatix, deduplication rules with compare-and-merge let you collapse duplicates during import rather than inheriting them. Capture geo-coordinates where you have them; where you do not, the field will backfill them during early visits.
- Items and the catalog. Load the item master with units of measure and their conversions (case, box, piece), barcodes where available, and the category structure reps will actually browse by. A catalog organized by your ERP's logic rather than a rep's mental model slows every order.
- Price lists and MRP. Load price lists per channel or region as they really operate, not as the price book claims. This is also the moment to write down scheme and discount logic explicitly — you will encode it as server-enforced rules later, and discovering in week 7 that nobody can state the scheme precisely is a common and avoidable stall.
What to watch qualitatively: can a supervisor look at the loaded outlet list for their territory and say "yes, that is my territory"? Until the answer is yes, do not proceed. Field teams forgive a plain app; they do not forgive wrong data about their own outlets.
Weeks 3–4: Beats and routes digitized, visit plans automated
With outlets in place, digitize how the field actually moves.
- Transcribe the paper beats first. Resist the urge to redesign territories while digitizing them. Capture today's beats as-is — even the irrational ones — so the field recognizes the plan as theirs. Territory redesign is a separate exercise you will be far better equipped for in month three, when you have real visit data.
- Build routes on the map. In xMatix, routes and beats are first-class objects built in a map-based Route Planner console, and route sequences can be optimized on real road networks rather than straight-line distance. Use optimization as a suggestion at this stage; let supervisors accept or override it so the plan keeps their local knowledge.
- Switch on automatic visit-plan generation. Once beats exist, visit plans should generate on schedule rather than being assembled by hand each week. xMatix auto-generates visit plans on a schedule and cleans up stale plans automatically. This is the moment the operation stops depending on someone remembering to make the plan.
What to watch qualitatively: run two weeks of generated plans past the supervisors who used to write them. Where the generated plan differs from what they would have written, one of two things is true — the master data is wrong, or the old plan was. Both are worth knowing before reps arrive.
Weeks 5–8: Rep rollout — pilot, champions, then the line
This is the phase where change management outweighs configuration.
- Start with a pilot team, not a pilot region. Pick eight to fifteen reps across different territory types — dense urban, thin rural, a van route — rather than one friendly region. You want the app to meet its hardest conditions while the group is still small enough to support personally.
- Recruit champions before the pilot starts. The most influential rep in each team, not the most senior. Give champions early access, take their complaints seriously and visibly, and let them do the peer training. A rollout announced by a champion lands differently than one announced by head office.
- Lead with what the app gives the rep. The offline-first design is your best opening argument: every screen works with zero signal, orders save locally and instantly, and nothing is lost when the network dies mid-visit — the outbox replays changes when connectivity returns. Show the catalog with barcode scanning, unit conversions, live scheme estimates, and frequently-ordered suggestions. Reps adopt tools that make their order faster; they resist tools that only make them observable.
- Sequence the features. Week one of the pilot: attendance check-in/check-out and the guided visit flow only. Week two: order capture. Then widen. Turning everything on at once produces shallow usage of everything; layering produces habits.
- Be explicit about tracking. Attendance uses a geo-stamp and optional selfie; route tracking runs with explicit user disclosure. Say so plainly, explain what supervisors will and will not see, and involve champions in that conversation. Trust lost here is expensive to rebuild.
- Roll out in waves. After two to three pilot weeks, expand team by team, with each new wave trained by the previous one's champions. Keep a fast feedback loop — layout and form changes in xMatix reach devices over the air without an app-store release, so fixes the pilot asks for can land while the pilot is still running, which does more for credibility than any kickoff meeting.
What to watch qualitatively: not adoption dashboards — conversations. Do reps describe the app as "how I take orders" or "extra work after I take orders"? Are supervisors referencing app data in their morning huddles unprompted? Is the paper duplicate dying on its own, or surviving in secret? Kill the paper process deliberately once — and only once — the field has stopped relying on it.
Weeks 9–12: Dispatch, optimization, and incentives
With reps live and data flowing, switch on the layers that need real data to be useful.
- Stand up the dispatch console. A live console is only as good as the data under it, which is why it comes last. The xMatix Allocation Console gives dispatchers Map, Gantt, and Board views with live rep positions, a KPI strip, and alerts — assigning visit plans, leads, cases, and service orders against a resource availability calendar, with trip playback for review. Train dispatchers on exceptions: the sick rep, the urgent complaint, the route running long. Routine days should run themselves off the auto-generated plans.
- Revisit routes with evidence. Now the territory redesign you deferred in week 3 has data behind it: actual travel patterns, visit durations, and productive versus non-productive outcomes. Re-run route-sequence optimization and consider beat changes with supervisors in the room.
- Switch on incentives last, and in the open. Configure incentive plans and targets per rep on the platform, computed from the same orders and visits the reps created. Run one cycle in shadow — computed and visible, not yet paid — so reps can challenge the numbers and finance can verify them before money moves. An incentive number reps can see mid-month, tied to records they recognize, is the strongest adoption lever in the whole rollout; a disputed first payout is the strongest sabotage.
- Close the loop with the back office. If van sales are in scope, this is when day-end reconciliation discipline (load, sell, take-in, reconcile) becomes routine, and field invoices flow into the same books that finance closes — one data model, no export-import seam to reconcile.
What to watch qualitatively: does the morning plan survive until evening without phone-call triage? When it does not, do dispatchers reach for the console or for their contact list? That single habit is the difference between owning a dispatch console and running live dispatch.
What this plan deliberately does not promise
No two operations start from the same data quality, geography, or team readiness, so treat every date above as a default to adjust, not a commitment to hold. Some phases compress — clean master data can halve phase one. Some stretch — a large van-sales fleet deserves its own pilot. The non-negotiable is the order: masters before routes, routes before reps, reps before dispatch, and incentives only when the underlying data has earned the field's trust. Get the sequence right, keep the feedback loop fast, and the ninety-day frame takes care of itself more often than not.