As-Is / To-Be
Compare the current and desired states using consistent measures, then separate the observed gap from possible causes and solutions.
Use this comparison when a team has many proposed actions but has not agreed on the problem. Describe the same subject, measure, and observation window on both sides. The three-part diagram is a working aid for gap analysis, not a causal model.
An observed current state, its evidence and measurement rules, and a desired state with a rationale and time horizon. Mark provisional targets explicitly.
A bounded gap statement and a next investigation or test. A proposed action remains separate from evidence that it will close the gap.
How to use it
- Define the subject and record the period, denominator, and evidence for the current state.
- Describe the desired state using the same definitions; state its timing and whether it is a requirement, benchmark, or hypothesis.
- Compare the two states. For rates, distinguish percentage-point differences from relative changes.
- List explanations still to verify and choose a proportionate next check before committing to a solution.
Keep its limits in mind
A target is not a forecast or a commitment. Different populations or observation windows may make a comparison invalid. The gap alone does not explain why it exists.
Define a repeat-purchase problem before choosing a tactic
Fictional retail example. A team is considering a follow-up process for first-time members. Its quarterly target is a test hypothesis, not a forecast.
On a small screen, scroll across the diagram to read it.
Measurement check: Confirm that every included member has completed 28 days of observation before comparing cohorts.
Interpretation: The change from 18% to 26% is 8 percentage points, or about 44.4% relative to 18%. Neither calculation establishes a cause.
Next step: Define the trial groups and measurement window before testing follow-up. Do not assume coupons or another tactic will produce the target.
Three reasons the apparent gap may be unusable
Use As-Is/To-Be when people are proposing actions before agreeing what needs to change. Name the same subject and measure on both sides, then describe the difference without inserting a cause. The desired state needs a rationale and an owner’s agreement; a diagram cannot create either.
Pause if the present state is an opinion without observable material. “Customers are disengaged” cannot be directly compared with a repeat-purchase target. Pause if the target uses a different population, event or window. A fourteen-day observation is not a completed twenty-eight-day result. Pause if the To-Be side is a preferred tactic, such as “send coupons,” rather than a condition the team wants to examine.
Keep three separate questions: what is observed, what is desired and what might explain the difference. The gap identifies a question worth investigating, not a diagnosis. A missing follow-up routine can coexist with low repeat purchase without being established as its cause. Avoid making an ambitious target sound feasible by writing it in the same visual style as a measured baseline.
Choose a gap, a goal, an explanation or a trial
The three-part comparison can lead to several kinds of work. Select the one supported by the current material rather than treating every framework as a compulsory next stage.
| Method | Material needed | Output |
|---|---|---|
| As-Is/To-Be | Comparable current and desired states | A bounded gap and open questions |
| SMART | A direction, baseline and constraints | A proposed measurable agreement |
| Logic tree | Several possible explanations or investigation paths | Branches and evidence checks |
| PDCA | An approved limited change and an observation plan | Recorded execution, outcomes and adjustment |
Worked case: a newer cohort has not finished its window
The earlier retail case reports an 18% baseline and a proposed 26% target, without supplying a baseline count. The following independent fictional ledger illustrates a measurement check; it does not recover that missing denominator. Cohort X1 contains 100 eligible first-time members, all observed for 28 complete days, with 18 making a qualifying repeat purchase. Cohort X2 contains 80 eligible first-time members, each observed for only 14 days so far, with 16 repeat purchasers recorded.
Eligibility and the repeat-purchase event are stipulated to be the same in this exercise. X1’s complete-window rate is 18/100 = 18%. X2’s 16/80 = 20% describes its fourteen-day observation, not its final twenty-eight-day rate. Do not call the two-percentage-point difference a twenty-eight-day improvement, count the remaining X2 members as final failures or omit them because they are unfinished.
For the proposed To-Be state, keep the same twenty-eight-day definition and label 26% as a hypothesis. Against X1, the target gap is eight percentage points, or about 44.4% relative to 18%. If a future comparable cohort had exactly 100 members, 26 repeat purchasers would correspond to that target; it is a conditional translation, not a forecast of eight additional repeat purchasers caused by follow-up.
| Record | Population / outcome | Window and status | Permitted interpretation |
|---|---|---|---|
| X1: As-Is illustration | 100 members / 18 repeat purchasers | 28 days complete | 18% under the supplied definition |
| X2: interim observation | 80 members / 16 repeat purchasers so far | 14 days only | 20% at day 14; final 28-day result unavailable |
| To-Be hypothesis | 26% proposed; no actual future cohort | Same 28-day measure intended | Target to discuss and test, not achieved |
| Gap against X1 | 26% minus 18% | Comparable definition required | 8 percentage points; cause unknown |
Conclude with a comparable measurement and an open cause question
The immediate conclusion is that X2 cannot yet answer the twenty-eight-day question. Confirm the eligibility rule, qualifying purchase, entry date and observation endpoint in the real records. Complete the window before interpreting its final rate. A fixed reporting deadline does not make an unfinished cohort mature, and changing the denominator to obtain a convenient number weakens the comparison.
Separately, propose investigating whether follow-up, product fit, availability or another explanation matters. These are hypotheses, not findings from the gap. Before a limited trial, agree the comparison, exposure, contact arrangements, resources and decision authority. Do not assume a coupon or any other tactic will produce 26%. Keep the target rationale and feasibility questions visible.
At review, report actual counts and complete windows alongside any changes in membership, offer or measurement. Preserve the original target and append revisions with reasons and actual agreement. Finishing a follow-up task, reaching a rate and demonstrating a causal effect are separate claims. If evidence is still missing, the useful output is a narrower next check rather than a manufactured success statement.
Check your work
- Both sides describe the same subject and metric.
- Each cohort has the observation time needed for that metric.
- The target has an explicit basis and is not presented as a guaranteed outcome.
- A numerical gap is separate from a proposed cause or intervention.
Framework references
References explain the framework and its scope; they do not validate the fictional example's results.
Put the framework to work
Copy the outline and add your question, evidence, and open questions.
Read a practical guide →Practice with an example
Switch denominators for the same cohort counts, distinguish the first task from later use, and leave with a bounded problem statement and evidence plan.
New users aren’t coming back. What should you check first? →