Change portfolio data entry
How much effort should change portfolio data take? A realistic answer to the data-entry question

Nov 18, 2026 | Change Measurement

Latest Articles

Join our newsletter!
Get the most insightful Change articles

In our conversations with heads of change and transformation at large regulated enterprises, the effort of data capture is the objection raised more often than any other, ahead of cost. One transformation lead told us that if a tool takes 80 per cent of people’s time to populate, it is self-defeating. A portfolio executive said their organisation already does project portfolio management and nobody is going to update another thing. Both are right. This article is for the heads of transformation and change, and the executives behind them, who have to judge whether change portfolio data is sustainable before committing anyone to it. The argument is that the effort is a design choice, not a fixed cost. You set it by working backwards from a handful of decisions, capturing a group of fields, reusing what project and HR systems already hold, and updating monthly. By the end you will have a minimum viable dataset, rules for granularity and cadence, a defensible effort budget per initiative, and a way to tell when you are collecting too much.

The data-entry objection is correct, and the usual answers are not

Every profession that has been asked to feed a system of record knows how this goes. In a 2024 survey of more than 1,200 US health professionals by the American Medical Informatics Association, reported by Healthcare Dive, nearly 75 per cent said documentation impedes patient care and more than 77 per cent said they finish late or work after hours because of it. Closer to the change function, Wellingtone’s State of Project Management Report 2026 found that 72 per cent of respondents spend half a day or more every month collating project reports.

The project managers in that figure are the people a change portfolio usually asks for updates. They already maintain a status report, a schedule, a risk log and a finance forecast. “Another thing to update” is an accurate description of the request.

What happens next is predictable. Contributors enter the minimum that stops the reminders. Defaults get accepted. Someone senior notices a go-live date that moved two months ago, and confidence in the whole dataset drops. The portfolio asked for more than its contributors could see a reason to give.

Three answers that do not gel with a busy project team

The first common answer is that the insight will be worth it. That may be true for the executive forum. It is not true for the person entering the data, who is the one investing the effort while the benefit goes to someone else later.

The second is to capture everything in case it proves useful. This is how datasets reach 60 fields per initiative, and this is what the incumbent is trying to avoid.

The third is that AI will do the entry. Part of this is real, and it is covered later. But a drafted record still needs a person to check it, as set out in our piece on whether you can trust an AI-drafted impact assessment or project update. Automation reduces typing. It does not reduce the number of judgements you have decided to collect. It also cannot dream up the correct data from nothing.

Work backwards from the decisions the portfolio has to support

Bart de Langhe and Stefano Puntoni made the general case in MIT Sloan Management Review: most analytics efforts start with the available data and go looking for a use, when they should start with the decision and work back. Their summary of decision-driven data analytics is to find data for a purpose instead of finding a purpose for data. Change portfolio data suits this approach, because the list of decisions a portfolio informs is short.

In a large enterprise, a change portfolio earns its keep on five recurring decisions:

  1. Whether to move, stagger or hold a go-live or other implementation activities because of what else is landing on the same people in the same weeks.
  2. Which business units are beyond their capacity to receive change in the coming two quarters, and what gets deferred or reshaped as a result.
  3. Where to place scarce change practitioners, when there is limited resource available.
  4. Whether a new initiative can enter the portfolio on the dates its business case assumes.
  5. What to tell the executive team this month: where the cumulative impact sits, what it puts at risk, and the so what.

We have written before about the decisions that change data makes better, so the point here is narrower. Every one of the five can be answered from the same small set of facts: which initiatives, what activites, hitting which groups, how hard, in which weeks, supported by whom.

A three-question test for every field

Before a field goes into the dataset, ask three questions. Which of the five decisions does it inform? Who makes that decision, and in which forum? What would they do differently if the value were higher or lower?

If you cannot answer all three, cut the field. A field such as “number of communications sent” usually fails the third question: no executive moves a go-live because an initiative sent nine emails instead of six. A field such as “impact level on claims assessors in March” passes all three. Run the test with your PMO counterpart in the room, because they will recognise which of your proposed fields they already hold.

The minimum viable dataset for change portfolio data

Applying that test across the five decisions leaves roughly ten fields per initiative. The table separates them by what matters most for effort: whether the value already exists somewhere or needs a fresh judgement.

FieldDecision it servesWhere it comes fromHow often it changes
Initiative name AllProject portfolio toolRarely
Accountable executive and project manager1, 4Project portfolio toolRarely
Delivery phase and status4, 5Project portfolio toolMonthly
Key dates: go-live, training, cutover1, 4Project scheduleWhen the plan moves
Impacted groups, picked from the organisation hierarchy1, 2HR system hierarchy, change lead selectsWhen scope moves
Headcount of each group2, 5HR system, never typedAutomatically
Impact windows: date ranges when each group feels the change1, 2Change lead’s judgementWhen the plan moves
Impact level per group per window, on a five-point scale1, 2, 5Change lead’s judgementWhen scope or approach moves
Change support assigned, in days per month3Change team resourcing sheetMonthly
Date confidence: firm, likely or tentative1, 4Project manager’s judgementMonthly

Three judgements and seven lookups

Only three fields need someone to assess: which groups are affected, when they feel it, and how much. Everything else is a lookup from a system that already holds the answer.

This changes the conversation with the portfolio executive who does not want another thing to update. The request is not “maintain a second portfolio record”. It is “give us read access to the one you have, and we will add additional insights you currently do not have”. Those judgements are also the part of the dataset most exposed to optimism and politics, a separate problem covered in our article on protecting the integrity of change impact data.

If you are starting from nothing, the three-tab structure in our guide to tracking multiple change initiatives holds this dataset for a small portfolio.

What to leave out until a decision asks for it

A second tier of data is legitimate but conditional: readiness and adoption measures, hours of training and briefing per group, customer-facing impacts, readiness and adoption assessment results. Each is valuable for specific initiatives at specific moments, and none is needed for every initiative every month. However, there is value in tracking readiness and adoption measures across the portfolio, you just need to ensure that your stakeholders are requesting this.

This tier-two data should be only worked on as requested, not by default. How granular is granular enough?

Granularity multiplies effort faster than field count does. Ten fields across 12 groups and 26 weeks is a different job from ten fields across three divisions and two quarters.

People: stop at the level where someone can act

Capture impact at the lowest level of the hierarchy at which a leader could make a different decision, and no lower. A chief operating officer can stagger two go-lives for the claims division. Nobody makes a portfolio decision about a team of nine, or about a named person.

In practice this means division, then business unit, then the large role groups inside a unit where load differs sharply, such as frontline agents, their team leaders and the specialists behind them. Depth can vary across the organisation: go one level deeper in units close to their limits and stay shallow where load is light. Every initiative touching the contact centre has to rate it at the same level (consistency of rating), or the numbers cannot be added up.

Time: enter date ranges, read weeks

Executives need to see load by week, because collisions happen in specific weeks. Nobody should type weekly values to achieve that. The person updating the record enters an impact window as a start date, an end date and a level. The tool, or a formula, spreads that across the weeks. One window per meaningful phase (readiness, training, go-live and early use, stabilisation) is enough, so four rows per group is a good approach.

Scale: five defined points beat a hundred undefined ones

Finer scales cost more to complete, because every rating becomes a small debate. They do not buy precision. In a study of more than 1,300 participants randomly assigned to response scales of two to eleven options, researchers at the University at Buffalo found no measurement advantage beyond six options. That research concerned personality questionnaires, not impact ratings, so treat it as a caution about false precision, not proof. In our experience a five-point impact scale with a written definition and an example for each point gives more consistent data, faster, than a ten-point or percentage scale with none. Our practitioner guide to change impact assessment covers how to define and calibrate the scale.

Initiatives: set an entry threshold

Not every change belongs in the portfolio as its own line. Set a threshold, for example any initiative that takes more than a day of a group’s time in a quarter, or that alters how a role is performed. Below the threshold, roll small operational changes into a single line per business unit, estimated once a quarter by the unit’s own operations lead. This one rule shortens the list considerably and loses little, because small changes rarely drive a portfolio decision on their own.

Update cadence and who does the updating

Why monthly is usually enough

Update cadence should follow decision cadence. Most enterprise portfolio and investment forums meet monthly, and the decisions listed earlier look one to three quarters ahead. A weekly refresh of data that feeds a monthly decision is wasted effort.

Monthly has one weakness: a plan can move materially between cycles. Handle that with triggers, not frequency. A change to a go-live date, a rescope, or a new impacted group prompts an update in the week it happens. We set out trigger points in our article on continuous change impact assessment, and the same triggers work for portfolio data. For most initiatives in most months, the correct update is “no change”, and it should take one click or one line.

One named updater per initiative

The model that lasts is one named updater per initiative: the change lead where one is assigned, and the project manager where there is none, working from a short guided update. The central team curates. It chases records that look stale, checks that units are rated at consistent depth, and prepares the forum view. Once a month, a representative of each business unit should also review the ratings for their area. That review is where the receiving side gets its say.

Enter once, reuse everywhere

European governments have adopted what they call the once-only principle: citizens and businesses provide a piece of information to public administrations once, and agencies reuse it instead of asking again. Apply the same principle to your contributors. If the project portfolio tool holds the go-live date, nobody types it again. If the HR system holds the headcount of retail lending, nobody estimates it.

No integration project is needed to start. A monthly export from the project tool and a quarterly export of the organisation hierarchy, imported into the change dataset, covers a lot of the ten fields. Whether the source is Microsoft Project, Jira or a PMO spreadsheet matters less than having one source for each field. Build the automated feed later, once the dataset has proved its value in the forum.

Setting an effort budget for change portfolio data

US federal agencies cannot ask the public for information without first estimating what it will cost the people answering. The government’s own guide to the Paperwork Reduction Act requires each agency to estimate the burden of a collection from the number of respondents, the frequency of response and the hours per response, counting time spent reading instructions, finding the data and reviewing it as well as typing. Few internal data requests face any equivalent test. A change portfolio should.

Build the budget in four steps:

  1. List every person who will contribute and what you are asking of them: set-up, monthly update, quarterly review, curation.
  2. Time each task with two or three real contributors. Do not estimate it from your own desk.
  3. Multiply by frequency and by the number of initiatives to get hours per month, and express it per initiative.
  4. Publish the figure to the contributors and the PMO, and commit to a ceiling.

A worked hypothetical: the effort budget for a 45-initiative portfolio

Take an insurer with 45 initiatives, with various business units.

Set-up is a one-off: 45 initiatives at about 5 minutes each, or roughly 4 hours over the first quarter, less where an existing assessment can be imported. In steady state, the monthly update takes 1-3 minutes per initiative (1.5 hours), central curation takes half a day (4 hours), preparing the forum view takes half a day if reporting is generated from the dataset (4 hours). That is less than 14 hours a month in total.

Signs you are over-collecting

Over-collection shows up as symptoms in the data and in the behaviour of contributors:

  • Fields where a third or more of the values are blank, defaulted or unchanged since set-up.
  • Fields that no report, forum paper or decision has drawn on in the last two cycles.
  • Arguments over whether an impact is a 3 or a 4, which means the scale is finer than anyone’s knowledge.
  • The central team retyping dates or headcounts that another system already holds.
  • Contributors keeping a private spreadsheet because the official record is too heavy to use as a working tool.

Run a field review once a quarter. For each field, record how complete it is, which decision used it last quarter, and roughly what it costs to maintain. Then keep it, move it to tier two, or delete it. Contributors notice when a field is deleted, and it earns cooperation.

The opposite failure exists too. If you cannot say which groups are hit, in which weeks, by which initiatives, you are under-collecting, and the three judgement fields are where to look.

Where a purpose-built platform removes effort, and where it does not

The design choices above apply wherever the data is stored. A platform changes how much of the remaining effort is typing. In The Change Compass, the tool that matters most to an effort budget is Change Automator, which sits beside Change Portfolio and Change Driver on the same platform.

Three capabilities bear directly on the effort question. The first is import: initiatives and impacts already held in spreadsheets can be brought across, so set-up does not mean retyping a year of work. The second is document upload. A practitioner uploads a project document(s) and the platform drafts a first impact assessment, which a person then reviews and approves, so the hour goes on correcting a draft and not on filling blank fields. The third is templates. With them, setting up a new initiative takes 2-3 minutes, and among users a monthly update cadence is typical. Once the data is in, the views built from it (the portfolio heatmap, the timeline and the Executive Dashboard) all come from that one record, so the monthly executive view does not have to be rebuilt by hand. It also then proactively identifies key risks and opportunities across the portfolio.

To find out more, book a demo.

Decide the effort first, then the dataset

Answer the data-entry question with a number. Set the effort budget for change portfolio data before you design the dataset, and let the budget discipline the design: five decisions, about ten fields, three of them real judgements, group-level granularity, date ranges instead of weekly values, a five-point scale, a monthly cycle with event triggers, one named updater per initiative, and every reusable field read from the system that already owns it.

Frequently asked questions

How much time should change portfolio data take to maintain?

As a planning guide, the named updater should need around 2-5 minutes per initiative per month, with most months recording no change. Central curation and forum preparation add one to two days a month for a portfolio of 40 to 50 initiatives. Higher figures usually point to too many fields or too fine a granularity.

What is the minimum data needed for a change portfolio view?

You need to know which initiatives hit which groups, how hard and in which weeks, plus who supports each initiative. That is about ten fields, seven of which can be read from project and HR systems. The other three are judgements: impacted groups, impact windows and impact level.

How often should change portfolio data be updated?

Monthly suits most large organisations, because portfolio and investment forums meet monthly and look one to three quarters ahead. Add an immediate update when a go-live date moves, scope changes or a new group is affected.

Who should be responsible for updating change portfolio data?

One named person per initiative, normally the change lead, or the project manager where no change lead is assigned. The central change team curates and quality-checks the data. Business unit representatives review the ratings for their area each quarter.

Can change portfolio data be pulled from existing project tools?

Most of it can. Initiative details, status and key dates sit in the project portfolio tool, and group structures and headcounts sit in the HR system. A monthly export and import is enough to begin with.

References