Home/Blog/Scorecards/How to Weight Scorecard Questions: Two L…
Scorecards

How to Weight Scorecard Questions: Two Levers Only

How to weight scorecard questions: use option values within a pillar and pillar weights between pillars, because question count never changes influence.

Cover of the FIKR Space guide: How to Weight Scorecard Questions: Two Levers Only

To weight scorecard questions, use only the two levers that actually work: the option values on a question, which set its scale within a pillar, and the pillar weights, which set the scale between pillars. Adding more questions to a pillar does not make that pillar count for more.

People weight scorecards by instinct. They add extra questions to the dimension they care about, or they carefully adjust weights until they sum to a round number. Neither does what they expect. This post explains how the arithmetic really behaves, so that the weights you choose are the weights the score uses. The grid itself is covered in how to choose scorecard pillars and bands.

How to weight scorecard questions: start with the option ladder

For a single-select question, the highest possible score is not something you declare. It is worked out as the highest value among the question's options. That single fact explains most weighting surprises, so it is worth stating plainly: the option ladder is your measurement scale.

Several consequences follow.

  • Adding a richer option changes everyone's percentage. Suppose a question tops out at one value and you add a new option above it. The maximum rises, so every earlier answer now represents a smaller share of the possible score. Old results and new results stop being comparable, and nothing recalculates the old ones.
  • Multi-select has no natural ceiling. Its maximum is the sum of every option. A question with several generous options can end up worth more than a whole pillar built on a small ladder. It is the one type whose weight grows with the number of options you happen to write.
  • A yes/no question can swallow a pillar. Put a two-option question scoring high next to neighbours whose ladder tops out at a small number, and it dominates the pillar. Author it as a two-option question on your own ladder instead.

The practical rule is to pick one ladder for the whole instrument and write it down, because nothing in the software will remember the choice for you. Two conventions are common, a short run of whole numbers and a run in steps of twenty-five. Either works if you stay with it.

Question type is a scoring contract

The kind of question you choose decides how it scores, not only how it looks. Four behaviours matter.

TypeWhat it asksHow its maximum is setBehaviour
Single-selectPick your levelHighest option valueMoves on automatically
Multi-selectHow many of these do you doSum of all optionsWaits for the respondent to press next
Rating scaleHow stronglyAlways one hundred, so it normalises itselfWaits, and starts with an answer already in place
Text inputTell meZeroDoes not affect the score

The text question is a deliberate exception. It adds nothing to the numerator or the denominator, so it drops out of the calculation rather than scoring zero. It is therefore the one question a respondent can safely leave blank. The rest of the types do count, and a rating scale begins with a midpoint position that registers as an answer even if the person never touches it. Keep that in mind when you read the results, as the post on finding dead questions in a scorecard explains.

Free guide, 39 pages: The Scorecard Playbook. Send me the guide.

Pillar weights are ratios

Pillar weights are divided by whatever they add up to. That means a set of weights written as equal twenties and a set written as equal ones describe the same scorecard. You do not need to force the weights to total one hundred, and you should stop doing the arithmetic to make them. Real instruments carry weights that total something odd and still score correctly.

What matters is the ratio between pillars. Two pillars with weights in the proportion two to one will count in that proportion, whatever the absolute numbers are.

Question count is not weight

This is the part that surprises people most. Within a pillar, scores are averaged first. A pillar with three questions and a pillar with twelve questions, given equal weights, count equally. Adding more questions to a pillar does not raise its influence on the final result.

So the instinct to add questions to the area you care about is a mistake if the goal is influence. It lengthens the instrument, which costs completions, and leaves the weighting unchanged. You are left with exactly two levers:

  1. Option values set the scale within a pillar.
  2. Pillar weights set the scale between pillars.

Making one question count double

In most engines a per-question weight field either does not exist or is stored and never read. If you want a question to matter twice as much inside its pillar, the honest way is to double its option values. A question that runs from zero up to one hundred in steps of twenty-five can be replaced by one that runs from zero up to two hundred in steps of fifty, sitting beside neighbours on the original ladder.

It is crude, but it is the only per-question weighting that this kind of engine honours. Use it sparingly, and write the reason down so the next person understands why one question has a different ladder from the rest.

When unequal pillar weights are justified

The default is equal weights. Use unequal ones only when you can defend the inequality out loud to someone who disagrees. A good test is to finish the sentence, "this pillar counts more because...". If the honest ending is "because I wrote more questions about it", the weight is following supply and not conviction. Weight should express an editorial view, not question supply.

SituationBetter choice
One pillar feels more importantRaise its weight, and be ready to explain why
One pillar has more questionsLeave the weights equal, since count does not matter
One question deserves more sayScale its option values, not its position
Weights do not sum to a round numberLeave them alone, the arithmetic normalises
FIKR Score, diagnostics that qualify: turn your audience questions into a quiz that finds buyers. See FIKR Score.

The denominator is a promise

Every question shown adds its maximum to the denominator, answered or not. A skipped question therefore counts as a zero and not as an exemption. That is correct only if the instrument insists on a complete submission, which is why careful engines refuse a partial one.

The corollary matters whenever you show some respondents fewer questions than others. Every path that narrows the question set must narrow the denominator with it. If a subset of pillars is shown but the score is still divided by the full instrument, the totals come out far too low and people land in the wrong band. Fixing only the numerator reproduces the same wrong number.

Freeze the scale before you launch

Because scores are written once, at completion, and never recalculated, every change to a ladder or a weight is a fork in the instrument. The discipline is straightforward:

  • Tune the model before you drive traffic.
  • Afterwards, batch your changes and date them in your own notes.
  • Never read a trend chart across an edit.

Editing the text of a question is safe. Editing the text of an option is not, because translations and in-flight respondents are tied to the option as it was. Treat any option change after launch as a new version of the whole scorecard. Once the scale is fixed, the next job is turning scores into bands, in how to set scorecard band boundaries.

Why a richer option moves everyone: a walk through

Take a question with four options worth nought, twenty-five, fifty and seventy-five. Its maximum is seventy-five, so a respondent who chose the top option scored the full amount, one hundred per cent of the question. Now add a fifth option worth one hundred because you decided the ladder needed a higher rung. The maximum becomes one hundred, and the same respondent's seventy-five is now three quarters of the question.

Nothing recomputes. Everyone who answered before is frozen at the old arithmetic, and everyone who answers afterwards is scored on the new. A trend line drawn across that edit shows a drop that no respondent caused. The cure is not clever: decide the full ladder first, with the top rung you will ever want, and only then let people answer.

Free guide, 39 pages: The Scorecard Playbook. Send me the guide.

A checklist for the measurement model

  • One option ladder is chosen for the whole instrument and written down.
  • No yes/no question sits on a different ladder from its neighbours.
  • Any multi-select question has been checked for how large its maximum becomes.
  • Pillar weights express a view I can defend, or they are all equal.
  • I know that adding questions to a pillar will not increase its influence.
  • Every path that shows fewer questions also shrinks the denominator.
  • Options are frozen before launch and before translation.

If every line can be ticked, the scale is ready. A wrong layout is only embarrassing. A wrong score is a false statement about someone's business, in a document they may act on, which is why this stage deserves the care.

FIKR Score, diagnostics that qualify: turn your audience questions into a quiz that finds buyers. See FIKR Score.

Frequently asked questions

Do my pillar weights have to add up to one hundred?

No. They are divided by their own total, so only the ratio between them matters. Equal weights of any size give the same result.

Will more questions make a pillar matter more?

No. Scores are averaged within the pillar first, so a short pillar and a long pillar with equal weights count equally. Use pillar weights to change influence.

Is it safe to add a better option to a live question?

It is not. The maximum is derived from the options, so a richer option lowers the percentage for every earlier answer, and old and new results can no longer be compared.

Which ladder should I choose?

Either of the common ones is fine. What matters is choosing one for the whole instrument, writing it down and keeping every question, including any yes/no, on it.

The full method, with the checklists, is in The Scorecard Playbook, a free 39-page playbook.

LG

Luis Goncalves

// Founder & CEO at FIKR Space

Three-time founder. Built and exited Evolution4All before this. Now building FIKR Space — the operating infrastructure underneath every innovation ecosystem (startups, accelerators, governments, investors). Lisbon-based, works global.

Claim your founding member spot.

Every founding member is hand-reviewed. Tell us who you are and how you want to co-create — we get back inside 48 hours.

Role
Job Title
Country
Cookies

We use cookies.

Essential ones keep the site working. The rest only run if you say so. No dark patterns, no pre-ticked boxes.Privacy policy

Co-create with us. It's about building it together.Co-Create With Us