Tool & guide

Top DST Sponsors: Missing-Data Treatment Explorer

By Sequoia ResearchPublished Read as Markdown

Missing-Data Treatment Explorer

Users identify missing fields and compare transparent treatments such as exclusion, separate disclosure, or withholding a score. The output shows how each policy affects coverage and ordering without silently converting unknown values to zero.

For an export, select Download in the tool, then Save file when it appears below. Your entries stay in your browser.

What Is This Article About?

This guide explains how missing evidence can change a sponsor comparison before anyone changes a score. The tool applies three clear policies to the same three profiles and four factors. It shows which profiles or factors are left out, how much of the original weight remains, and when no ranking is possible. A fictional example makes the tradeoffs visible without treating an unknown fact as a zero.

Table of Contents

What Missing Data Means in DST Sponsor Research

A missing field means the entered record does not meet the rules for that cell. It does not mean that the sponsor failed, that a fee is zero, or that a report does not exist. The source may be private, unavailable to the reviewer, out of date, or simply not entered yet. Preserve the reason in the notes where you know it.

DST means Delaware statutory trust here. This tool helps organize a comparison of sponsor evidence. It does not review a particular offering, choose a sponsor, fetch private documents, or rate the risk of an investment. You supply all facts, scores, labels, and source checks.

The U.S. Securities and Exchange Commission, or SEC, explains that private placements can provide less public information than registered offerings. That makes the scope of a research file important. A public-page review and a review of private offering records may have very different gaps. Neither should be described as the other. SEC private-placement bulletin

The explorer keeps the missing-data choice visible. One policy removes incomplete profiles from scoring. Another removes factors that lack support for any profile. The third reports evidence and gaps without an overall score. Each choice answers a different question. The software shows their effects; it does not select the right research policy for you.

Set the Profiles, Factors, and Evidence Rules

Enter exactly three profiles. Give each a distinct name and a clear scope, such as the legal entity, product type, and period under review. Two rows with nearly identical names can lead to confused results. The tool rejects names that match after case, spacing, and Unicode normalization. It does not resolve real corporate identities for you.

Define four factors with scores that run from zero to 100. Explain what the endpoints mean and why a higher score is more favorable under your rubric. A rubric is your written scoring rule. Do not mix raw years, dollars, and percentages in these fields without explaining how you convert them to a common scale.

Each factor also needs a weight from zero to 100 and a reason for that weight. Weights may use two decimal places and need not total 100. At least one must be positive. A factor with zero weight is inactive: its records remain in the file, but it does not affect coverage or either scored policy.

For each profile and factor, enter a score, source reference, page or other location, and evidence date. Then confirm that you checked the source against the profile, cutoff, and rubric. Notes can explain limits but are optional. The manual check records your action; it does not make the software a source verifier.

A cell is supported only when it has all required fields and a checked confirmation. A score alone is insufficient. A source reference without a score is also incomplete. The output lists each missing part so another reader can see why a profile or factor was excluded.

Set an evidence cutoff no later than the current date in Coordinated Universal Time, or UTC. Evidence dates cannot follow that cutoff. Dates must be real calendar dates from 1900 through 2200. The cutoff limits the dates you enter; it does not establish that a document is the latest available version.

Policy One: Withhold Incomplete Profiles

This policy preserves the full set of positive-weight factors. A profile receives an aggregate score only if every active cell is supported. A profile with one missing required field has its whole score withheld. Its inputs and gaps remain in the output and both downloads.

The weighted score uses each factor's share of the total weight. Multiply each score by that share, then add the products. All eligible profiles use the same factors and denominator. This makes their arithmetic comparable within the defined file, though it does not prove the rubric itself is fair or useful.

At least two complete profiles are needed for ranks and a first-place set. If only one is complete, the tool shows its score but withholds ranks. A lone complete row is not automatically declared the best. If none is complete, all scores and ranks are withheld.

The benefit is a stable research question. You do not drop a factor just to produce a result. The cost is a smaller group. If records are easier to obtain from certain firms, the remaining profiles may reflect access rather than the full sponsor market. The tool cannot correct that selection problem.

Read the excluded profiles before focusing on the scored rows. A missing fee schedule and an unchecked page location may both block a score, but they call for different research work. The gap list helps you decide what to request or inspect next. It is not a list of proven deficiencies at the sponsor.

Policy Two: Exclude Incomplete Factors

This policy keeps all three profiles but changes the factor set. A positive-weight factor is included only when all three profiles have supported cells for it. If one profile lacks support, that factor is excluded for everyone. The remaining weights are scaled to total 100% within the reduced comparison.

The output names the included and excluded active factors. It also reports the share of original weight retained. If factors worth 60 out of an original 100 remain, the retained share is 60%. Their new normalized weights add to 100%, apart from display rounding. Those are two different percentages with different meanings.

When at least one factor remains, all three profiles can be scored on that common set. When none remains, every score and rank is withheld. The tool does not fall back to zero scores or invent a factor to keep the table populated.

This method can make an incomplete file easier to compare, but it changes the question. If you exclude a fee factor, the new score no longer reflects that factor. A higher result after exclusion does not show that fees improved. It shows that the calculation stopped using the fee evidence.

A factor may matter too much to drop. If the research question requires it, the proper response may be to gather more evidence or leave the ranking unfinished. The explorer deliberately does not make that judgment. Use the retained-weight and excluded-factor labels to make the decision clear to anyone reading the output.

Policy Three: Keep Separate Disclosures Without Scores

The disclosure-only policy supplies no aggregate score, rank, or leader. It preserves the entered factor scores, source details, gaps, and coverage measures. This is a deliberate output, not a calculation failure. A research file can be useful before it supports an overall comparison.

Use this view when the factors should remain separate or when the gaps are too important to hide behind a reduced model. You can explain that one source is complete while another remains open. You do not need to combine every observation into one number to identify the next research step.

A user-created score also differs from a credit rating. The SEC describes credit ratings as opinions about creditworthiness, with limits on the risks they address. The explorer is not a credit-rating service, a recommendation, or a forecast of loss or return. SEC credit-ratings bulletin

The three policies are shown together so you can discuss the tradeoff. There is no cross-policy rank-change number. A profile may be absent from one ranking while all profiles appear in another. Subtracting those positions would imply that the underlying comparisons were the same when they were not.

Work Through the Fictional Missing-Data Example

Load the fictional example to see a complete set of form fields. Cases A, B, and C are invented profiles, not actual sponsors. All sources and scores are examples. The blank cells are intentional research gaps.

Factor Case A Case B Case C Entered weight
Record coverage 90 60 70 50
Method clarity 50 Unknown 70 20
Sector relevance 60 80 Unknown 20
Reporting detail 80 50 70 10

Case A has all four active cells supported. Cases B and C each have three. Because each missing factor carries weight 20, B and C each have 80% supported-weight coverage. A has 100%. These percentages describe entered evidence coverage under the example rules, not confidence, quality, or safety.

Under the complete-profile policy, only A receives a score. Its contributions are 45 from record coverage, 10 from method clarity, 12 from sector relevance, and eight from reporting detail. The sum is 75. B and C are withheld. There is only one eligible profile, so no rank or first-place set is reported.

Under the shared-factor policy, method clarity is excluded because B lacks support. Sector relevance is excluded because C lacks support. Record coverage and reporting detail remain. Their original weights total 60, so the retained share is 60% of the original 100.

The new weights are 50 divided by 60 and 10 divided by 60. They display as 83.33% and 16.67%. A's score is (90 × 50 + 80 × 10) ÷ 60, or 88.33 when rounded. B's is (60 × 50 + 50 × 10) ÷ 60, or 58.33. C's is 70.00 because both of its remaining scores are 70.

The shared-factor ordering is A first, C second, and B third. This does not contradict the earlier result. The earlier policy could not rank a group of one. The shared-factor policy ranks three profiles on fewer factors. A's increase from 75 to 88.33 reflects that changed formula, not improved evidence or better sponsor performance.

For another exercise, leave B's method score blank but restore C's sector score to 70 with complete supporting fields and a checked confirmation. The shared set then includes record coverage, sector relevance, and reporting detail. Its original weight totals 80. The scores become 81.25 for A, 63.75 for B, and 70.00 for C. A and C also become eligible under the full-factor policy, with scores 75 and 70.

In the disclosure-only view, both versions still have no aggregate scores or ranks. The evidence changes remain visible without forcing an overall judgment. Save the versions separately if you want to explain how the extra support changed the other two policies.

Read Coverage, Zeros, and Ties Correctly

Coverage has two forms. The supported-factor count tells you how many active factors meet the evidence rules. Supported-weight coverage tells you how much of the selected weight those factors carry. Three of four factors is 75% by count, but it may represent 80% by weight, as in the example.

A zero score is different from a blank. If zero is supported under the stated rubric, it enters the weighted calculation and still counts toward evidence coverage. A blank score is unknown and fails the cell's support rule. Replacing blanks with zeros would add an assumption this tool never makes.

Weights of zero are different again. They switch off a factor for the current calculation. A disabled factor's incomplete fields remain in the saved evidence, but they do not block a profile or reduce active-weight coverage. Giving that factor positive weight makes its evidence relevant again. Rebuild the results and inspect the effect.

Ranks use exact products of entered hundredth units before display rounding. Equal exact scores share a rank. The rank is one plus the number of eligible profiles with a strictly higher score, so two tied leaders followed by a lower score receive ranks one, one, and three. No alphabetical tie-breaker is applied.

Displayed scores use two decimal places. Two different exact scores may look equal after rounding while retaining different ranks. This is a display limit, not a hidden judgment. The file retains the arithmetic inputs and numerators so the result can be traced.

Use the Results to Plan Further Research

Start with the gaps that change the policy result most. If one missing location blocks a fully supported profile, inspect that reference. If a missing material factor causes the shared-factor policy to narrow the question too far, keep that issue open. Ease of completing a field should not determine its importance.

The Financial Industry Regulatory Authority, or FINRA, discusses covered firms' duties to investigate private placements and assess supporting information critically. A set of filled boxes does not replace that work. This tool organizes evidence claims and calculations; it does not certify that a broker's or investor's review is sufficient. FINRA Regulatory Notice 23-08

Missingness itself may follow a pattern. A source may be available only for larger firms or for completed programs. The resulting comparison could omit a different part of the market. The software does not estimate what absent facts would have been. Record the access limits and avoid calling the remaining sample representative without evidence.

Edit, Export, and Preserve the Full File

Every form edit removes the old results and disables downloads until you compare again. Changing a score, source detail, or note clears that cell's manual check. Changing a profile name or scope clears its four checks. Changing a factor name or rubric clears that factor's checks across all profiles. A cutoff change clears every check.

Weight and weight-reason edits retain the source checks, because they change the policy inputs rather than the score evidence. You still must rebuild. Invalid numbers, impossible dates, or dates beyond the cutoff produce an error and preserve the form for correction. Negative scores, scientific notation, and values with more than two decimal places are rejected.

Both downloads preserve all three profiles, all four factors, source fields, notes, missing-field lists, and every policy result. The JSON file holds structured data; the text file is a readable copy. The full JSON has a 500,000-byte UTF-8 limit, with separate field-length limits. Neither output is an abbreviated list of winners.

There is no import, autosave, account storage, or source lookup. Save before leaving the page. Loading the example or clearing the form requires an in-page confirmation. Cancel keeps the form, and new typing cancels an older pending confirmation. A later click cannot apply that stale reset to newly edited entries.

FAQ

Does the Tool Fill Missing Scores With Zero?

No. A blank score remains unknown. Unsupported active cells can withhold a profile or exclude a factor, depending on the policy. An intentional zero is accepted only with the same required evidence fields and manual check as any other score.

Which Missing-Data Policy Is Best?

The tool does not choose one. Withholding profiles preserves the factors but narrows the group. Excluding factors preserves the group but narrows the question. Separate disclosure avoids an overall score. Choose based on the research purpose and explain the tradeoff.

Why Does One Missing Field Affect Other Profiles?

The shared-factor policy requires a common supported set. If one profile lacks support for a factor, that factor is excluded for everyone in that policy. This prevents different profiles from being averaged over different factors within the same ranking.

Does Full Coverage Mean the Research Is Correct?

No. Coverage means the entered fields and checks meet the tool's rules. A source can be misunderstood, a rubric can be weak, and a comparison can be poorly designed. The software cannot detect those errors merely because every field is filled.

Can I Use Different Numbers of Profiles or Factors?

This version supports exactly three profiles and four factors. You can make a factor inactive by assigning zero weight, but its definition and records remain. It does not merge files or create a larger ranking from several separate runs.

Why Is a Complete Profile Sometimes Unranked?

The complete-profile policy requires at least two eligible profiles for ranking. A single complete profile can have a weighted score, but there is no meaningful comparison group within that policy. Its rank and the first-place set are therefore withheld.

About Sequoia Research

Sequoia Research, LLC provides independent research on DSTs and other real estate investment offerings. This educational library helps readers inspect evidence, explain assumptions, and keep open questions visible when comparing investment documents and providers.

Sources