Automation introduced: what happens in the first six months
The course of an introduction is remarkably predictable: first momentum, then a dip, then an unspectacular phase in which it actually starts to work. Know the dip and you can sit through it.
In short
- In month two everything takes longer than before — that is the normal state, not a sign of a wrong decision.
- Most introductions fail at one of three points: uncleaned data, too many processes at once, no owner after the setup.
- The first dependable signs come in month four — and in the inflow figures, not in closes.
- After six months three to four processes should be running, not ten.
The course
Month 1 — tidying up
The month in which nothing visible happens: clean the data, merge duplicates, settle the fields, set up access. One to three days of pure tidying.
Feeling: slow going. Danger: cutting it short. Import dirty data and you have an expensive dirty system afterwards.
Month 2 — the dip
The first process is running, and everything takes longer than before. Every action is new, every special case has to be resolved. The question "was that a good idea?" arrives reliably.
Feeling: worse than before the introduction. Danger: stopping.
Month 3 — getting used to it
The first process runs without thinking, the second gets added. The first special cases are resolved, a routine emerges.
Feeling: neutral. Danger: adding too much at once.
Month 4 — first signs
The inflow figures move: more enquiries answered, shorter response time, fewer forgotten follow-ups. Not closes yet — it is too early for those.
Feeling: cautiously positive. Danger: looking at closes too early.
Month 5 — sharpening up
The first review: which fields stay empty, which processes run into nothing, which message never gets opened? Tidying rather than extending.
Feeling: productive. Danger: skipping this month.
Month 6 — viability
Three to four processes run reliably, the figures are dependable, and the system gets used even when nobody reminds anyone.
Feeling: normal — and that is the goal.
Worth knowing
The dip in month two is so reliable that it belongs in the plan — not as a risk but as the expected course.
Fail to announce it and it gets experienced as evidence that the decision was wrong. Announce it and it gets experienced as confirmation that everything is going to plan. The same month, two completely different conclusions — and one leads to abandonment, the other to month four.
The three points where it fails
| Point | When | Warning sign | Remedy |
|---|---|---|---|
| Uncleaned data | month 1 | "we will tidy that up later" | a shutdown date for the old source |
| Too many processes | month 3 | five started at once | one per quarter |
| No owner | month 5 | nobody does the review | an appointment in the calendar, with a name |
What gets measured when
- From month 1: the share of enquiries with a documented first contact, and the time to it. Reacts immediately.
- From month 3: the number of qualified enquiries and forgotten follow-ups. Reacts within weeks.
- From month 6: the close rate per stage. Before that there are too few completed cases.
- From month 12: the cost per customer won. With longer cycles, correspondingly later.
The month deciding the next two years is month five — the review. It looks like a step backwards, because nothing new gets built, and so it regularly gets skipped.
Yet it is the point at which a setup becomes a system: fields that always stay empty get removed; processes nobody uses get switched off; messages nobody opens get replaced. Skip those two hours and two years later you are running a system in which half the fields are empty and nobody knows why they are there.
What is realistic after six months
| Realistic | Not realistic |
|---|---|
| 3 to 4 processes running reliably | ten processes, all maintained |
| Response time considerably shortened | twice as many closes |
| No more forgotten follow-ups | fully automatic sales |
| Dependable inflow figures | dependable cost per customer |
| 2 to 6 hours of maintenance a month | a system that runs with no maintenance |
Help me plan the introduction of marketing automation over six months – realistically, including the expected dip. Our situation: - People in customer contact: [number] - Enquiries per month: [number] - Typical time to close: [details] - State of our contact data: [clean / grown over time / unclear] - Who would be the owner after the setup: [person or "nobody"] - Available time per week: [hours] Tasks: 1. Draft a month-by-month plan for six months with tasks, effort and what is to be expected in each month – including the dip in month two. 2. Name which process we should set up first, and justify it from our situation. 3. Settle which metric gets looked at from which month – and which ones we expressly do not look at until then. 4. If no owner is named: say clearly what that means for success. 5. Phrase a review checklist for month 5. 6. Name what should realistically be achieved after six months and what should not. Do not recommend specific products.
In closing
The course is predictable, and that is the most useful information of all: month one is slow going, month two feels like a step backwards, and from month four something moves.
Announce the dip as part of the plan and you sit through it. And actually do the review in month five and after two years you have a system rather than a setup.
Common questions
How long does introducing marketing automation take?
Around six months to a viable state: one month of tidying up, one month with a noticeable slowdown, one month of getting used to it, first measurable signs from the fourth, the review in the fifth, and normal running in the sixth with three to four processes going.
Why does everything take longer in month two?
Because every action is new and every special case has to be resolved for the first time. That is the normal state, not a sign of a wrong decision — announce it in advance and it feels on plan; leave it unannounced and it feels like evidence of a mistake, and people stop.
What do introductions fail on?
Three points: uncleaned data in month one, too many processes started at once in month three, and a missing owner from month five. The third is the quietest — a system with no owner does not fail, it goes stale, and the damage only stands out when a customer is puzzled.
When do you see the first results?
From month four, and in the inflow figures: more enquiries answered, shorter response time, no forgotten follow-ups. Close rates become informative from month six at the earliest, cost per customer won from month twelve — with longer cycles, correspondingly later.
What is realistic after six months?
Three to four reliably running processes, a considerably shortened response time, no forgotten follow-ups and dependable inflow figures — at two to six hours of maintenance a month. Not realistic: ten maintained processes, double the closes, or a system that runs with no maintenance.
Marketing that sets itself up
The Studio Engine beta is live. Claim your spot and help shape it from the start.
Join the beta →