Pure Benchmarks · Guide

How to Track Your Investment Decisions and Know Which Ones Were Best

Short answer

There are two methods and they answer different questions. A decision journal, written before you buy, records the thesis, the time horizon, what would prove you wrong and what you would have bought instead; it is the only way to judge decision quality rather than outcome, but it only works going forward and most people never start one. The second method works backwards: reconstruct what the portfolio would be worth today had you made no changes at all, then measure every buy, sell and hold against that do-nothing baseline. That is decision-level attribution, and it needs only your transaction history, which already exists.

Tracking what you owned is easy and almost every broker does it. Tracking the decisions you made is a different problem, and it is the one worth solving, because a stock can rise 30 percent on a market rally while the thesis you bought it on was completely wrong. Judging a decision by whether it made money is outcome bias, and hindsight makes every past decision look obvious. This page covers both ways to separate decision quality from luck: the journal you keep going forward, and the counterfactual you can rebuild from transactions you have already made.

What a decision journal records

Before each purchase, write down why you are buying it, what you expect to happen, over what period, what would prove the thesis wrong, how much you are willing to lose, and what you would have bought instead. That last line is the one most people skip and the only one that makes the entry measurable later, because it supplies the reasonable alternative you will eventually compare the outcome against. When the position closes, record the return, the result relative to an appropriate benchmark, and a separate verdict on whether the decision worked as distinct from whether the stock went up.

Why journals fail in practice

A journal requires you to have started one before the decision, which means it is worthless for everything you already own. It also relies on the discipline to write an honest entry at the moment you are most excited about a purchase, and on remembering to close the loop months later. Most journals are abandoned inside a quarter. The method is sound; the failure rate is the problem, and it is why the second method matters.

The counterfactual method: measure against doing nothing

Your transaction history already contains every decision you made. Freeze the portfolio as it stood before each change, price that frozen version forward to today with real historical prices, and you have a do-nothing portfolio: what you would hold now if you had left it alone. Every buy, sell and rebalance is then measurable as the gap between what actually happened and what doing nothing would have produced. No journal required, no memory required, and it works retroactively on decisions made years ago.

Where the return number misleads you

A portfolio-level return blends decision quality with market direction and with the timing and size of your deposits. Money-weighted return fixes the deposit distortion and is what Sharesight reports, which makes it a genuine improvement over a simple percentage. It still measures the portfolio rather than the decisions inside it. Contribution and attribution analysis, the kind Koyfin produces, goes one level deeper by splitting the return across holdings and sectors, but its baseline is a model portfolio or an index, not the version of your portfolio where you did nothing.

What each tool actually tracks

Sharesight tracks trades and reports money-weighted performance with contribution analysis, which answers which holdings drove the return. Koyfin produces allocation and selection attribution against a model portfolio, built for analysts. Morningstar Investor scores holdings rather than decisions. Capitally and Stock Rover both evaluate trades against an index. Pure Benchmarks, our own product, builds the do-nothing counterfactual from connected holdings and scores each decision against it; it does not do tax reporting or model-portfolio attribution, and those gaps are marked in the same table as everyone else’s.

How this comparison works

Which of these tools attributes a result to the decisions you actually made, and which ones attribute it to sectors, to holdings, or to nothing at all?

Decision journals and attribution tools are both marketed as decision tracking and they are not the same thing. A journal records your reasoning before the trade. Attribution measures the outcome afterwards. This page compares attribution tools and marks the journal capability as its own row.
The second question is who is allowed to buy it. Several of the strongest attribution engines in this category are sold to advisors or to institutions, and the attribution module specifically sits above a retail price point.
The third question is what the result is measured against. Every tool here can measure against an index. Only one measures against other real portfolios as well.
Capabilities and prices were read from each vendor's own published pages on the date shown. Vendors change both.
Pure Benchmarks is our own product. It does not produce a Brinson decomposition and it has no journal, and both gaps are marked in the table rather than described around.
FactSet, Bloomberg PORT, Confluence StatPro and Zephyr also serve this category and are institutional products sold through direct sales; they are outside the scope of a page about tools an individual can actually buy.

The comparison

Capabilities verified against each vendor's published pages on September 17, 2026.

Capability	Pure Benchmarks	Koyfin	MSCI (Barra)	IBviz	Morningstar Investor	Sharesight	Capitally	Stock Rover
Decision attribution	Yes	No	No	No	No	No	Yes	Partly
Allocation vs selection	No	Yes	Yes	Partly	Partly	Partly	No	No
Your real portfolio	Yes	Partly	Partly	Yes	Partly	Yes	Yes	Yes
Self-serve retail	Yes	Partly	No	Yes	Yes	Yes	Yes	Yes
Any broker	Yes	Partly	Partly	No	Partly	Yes	Yes	Yes
Peer baseline	Yes	No	No	No	No	No	No	No
Advisor review	Yes	Partly	No	No	No	No	Partly	No
Decision journal	No	No	No	No	No	No	No	No
Free real output	Yes	No	No	Partly	No	Partly	No	No
Price	Free account	Free plan, attribution from $209 per month	Enterprise, price on request	Free tier, Pro at $15 per month	Subscription, free trial	Free plan up to 10 holdings, paid plans above that	Trial only, no free plan	Premium at $17.99 per month

✓ yes · ~ partly · ✗ no

What each capability means

Attributes results to the individual changes you made, not only to holdings or sectors

Classic attribution answers which holding or sector drove the return. Decision attribution answers a narrower and more uncomfortable question: did your trading decisions beat doing nothing? It is a buy-and-hold counterfactual — the tool freezes the portfolio as it stood before the change, runs that do-nothing portfolio forward alongside the real one, and reports the dollar difference. Those are different calculations and most tools only do the first.

Produces the classic allocation, selection and interaction decomposition

The institutional standard, usually called Brinson attribution. It separates the return you earned by weighting a group differently from the benchmark from the return you earned by picking better holdings inside that group.

Runs on your own linked accounts rather than a model or a manually built portfolio

Attribution on a hypothetical portfolio is an exercise. Attribution on the account you actually own, with the trades you actually placed, is a review.

An individual can sign up and use it without an advisor licence or an institutional contract

Much of this category is sold to advisors and asset managers. If attribution sits behind an advisor tier or a sales call, it is not available to the person whose money it is.

Works across brokerages instead of being tied to one

A tool locked to a single broker can only attribute the part of your money that sits there, which for most people is not the whole picture.

Compares the outcome against other real investors, not only against an index

An index is one baseline. What comparable investors did with the same market over the same days is a second one, and it is the one that tells you whether the decision was unusual or ordinary.

Can be used to review decisions somebody else made on your behalf

If an advisor or a managed account is making the changes, attribution is the only way to see whether those changes earned their fee.

Records your reasoning before the decision, as a journal

Journalling captures intent in advance; attribution measures outcome afterwards. They are routinely confused because both are called decision tracking, and almost no tool does both well.

Free tier produces real attribution output rather than a trial countdown

Several tools here have a free plan where attribution specifically is the thing the free plan does not include.

Every tool in detail

Pure Benchmarks

A buy-and-hold counterfactual for every change you made: each decision is scored against the do-nothing portfolio you would still hold if you had left it alone.

Best for: Individual investors reviewing their own decisions, or an advisor's, without an advisor-tier subscription.

Pricing: A free account produces the decision report on a linked portfolio. Attribution is not held back for a paid tier.

Strengths

Freezes the portfolio as it stood immediately before each change and runs that buy-and-hold version forward alongside the real one, so every trading decision carries its own dollar figure against doing nothing
Adds a second baseline beyond the index: verified connected portfolios in the same standardized risk category
Applies identically to changes an advisor or workplace plan made on your behalf
Read-only by design and unable to move money in any linked account

Limitations

Does not produce the classic Brinson allocation, selection and interaction decomposition; the unit of analysis is the decision rather than the sector weight
No decision journal. It measures what a change did, not what you expected it to do
Requires a read-only brokerage connection; there is no manual-entry-only portfolio
Does not track real estate, private holdings or self-custody crypto wallets

Koyfin

The most complete Brinson attribution available outside an institutional contract, sold on advisor plans.

Best for: Advisors, and individuals willing to pay an advisor price, who want allocation and selection effects done properly.

Pricing: Free, Plus at $39 per month and Premium at $79 per month exist, but Model Portfolios and Client Portfolios, which is where Attribution lives, start at Advisor Core at $209 per month and Advisor Pro at $299 per month.

Strengths

Separates performance against the benchmark into allocation, selection and interaction effects
Factors in every rebalance during the period, using the portfolio's weights throughout rather than only the weights at the end
Any equity ETF can be the benchmark, and it uses that ETF's historical constituents rather than only its current holdings

Limitations

Attribution sits on the advisor tiers, so the entry price for the feature is $209 per month
Attributes to groups and holdings, not to the individual changes you chose to make
Custodian integration is an advisor-plan feature, so an individual is often working from a model portfolio rather than a live linked account

MSCI (Barra)

The institutional reference implementation, and not purchasable by an individual.

Best for: Asset managers, asset owners and performance teams.

Pricing: No published price and no self-serve signup. The published route is "Request a demo" or "Get in touch".

Strengths

Brinson, fixed income, factor-based and multi-portfolio attribution models in one system
Proprietary framework for detailed curve and spread attribution on fixed income
The methodology much of the rest of the category is measured against

Limitations

Sold to "asset managers, asset owners and performance teams", not to individuals
No published price and no self-serve access point
Attribution is to factors and groups, not to the decisions a private investor made

IBviz

Deep attribution visualizations at a retail price, for one broker only.

Best for: Interactive Brokers clients who want contribution and risk analytics without an advisor subscription.

Pricing: Lite is free with 80+ visualizations on a 6-month lookback. Pro is $15 per month for the full analytics and extended lookback. Corporate is $350 per month.

Strengths

Shows which assets drove returns, the FX impact on profit and loss, and risk concentration
A genuinely free tier with the full visualization set on a six-month lookback
The cheapest real attribution on this page by a wide margin

Limitations

Interactive Brokers accounts only, so anything held elsewhere is invisible to it
Attributes to assets and asset classes rather than to the decisions you made
No comparison against other investors at all

Morningstar Investor

Holdings analysis rather than attribution, and the most widely recognised name a retail investor will meet first.

Best for: Fund-heavy investors who want overlap, style and fee analysis against a chosen benchmark.

Pricing: Sold as a subscription with a free trial. No price is published on the product page, so none is stated here.

Strengths

Portfolio X-Ray breaks down allocation, sector weightings, style and fees across funds and stocks together
Benchmarks can be chosen across a range from aggressive to conservative, and indexes added for comparison
Backed by Morningstar's own fund research and ratings

Limitations

Analyzes what you hold rather than what your changes did
No decomposition of a result into the decisions that produced it
A trial rather than a free tier

Sharesight

The strongest record of what happened, which is the input attribution needs, rather than attribution itself.

Best for: Investors who want accurate multi-broker performance, dividend and tax records with a benchmark attached.

Pricing: A free plan tracks up to 10 holdings with no credit card. Benchmarking is listed on the Starter plan and above.

Strengths

Separates capital gains, dividends and currency movement, which is the decomposition most retail tools skip
Benchmarks against any of more than 750,000 supported stocks, ETFs and funds
Excellent multi-broker and multi-currency record keeping

Limitations

Contribution and benchmark comparison rather than attribution to decisions
Benchmarking sits on the paid plans
No comparison against other investors

Capitally

The closest thing to a buy-and-hold counterfactual outside our own product, benchmarking your live portfolio against a frozen version of itself.

Best for: Investors who will maintain their own data by hand in exchange for keeping it out of an aggregator entirely.

Pricing: A 14-day trial with no credit card, then paid Sailor, Navigator and Captain tiers. Sailor caps at 1 project and 50 assets; Navigator lifts the asset cap; Captain adds stock options and private equity.

Strengths

Runs up to ten simultaneous benchmarks, one of which can be a buy-and-hold version of your own portfolio, which is the counterfactual most tools never build
Splits every return figure into capital gain, dividend income, currency movement, fees and tax rather than reporting one opaque percentage
Supports time-weighted, money-weighted and IRR side by side, so contributions do not masquerade as performance
Never connects through Plaid, Yodlee or any aggregator; data stays out of a third party entirely

Limitations

No broker connection of any kind. Data arrives by CSV or Excel upload, spreadsheet paste or manual entry, with presets per broker, and the vendor states plainly there is no background auto-sync
The comparison target is an index, fund or your own frozen portfolio, never other real investors
No permanent free tier; the 14-day trial is the only unpaid access
Entry tier caps at 50 assets and a single project

Stock Rover

A research and screening platform whose Trade Evaluator scores individual trades against an index rather than against the portfolio you gave up.

Best for: Research-led investors who want deep screening and are content measuring trades against a chosen benchmark.

Pricing: Essentials is $7.99 per month, Premium $17.99 and Premium Plus $27.99, with annual and two-year billing reducing the rate. The Trade Evaluator specifically requires Premium or above.

Strengths

The Trade Evaluator reports whether each trade added to or subtracted from performance, counting dividends gained or lost along the way
Scores at both the position level and the individual trade level rather than only the whole portfolio
Trades can be measured on an absolute basis or against an index, ETF or fund you choose
Sits inside a research platform with several hundred metrics and ranked screening

Limitations

The baseline is an index or fund you nominate, not the portfolio you would still hold had you not traded, so it answers a benchmark question rather than a counterfactual one
The Trade Evaluator is gated behind Premium at $17.99 per month; the free tier does not include it
No comparison against other real investors
No allocation and selection decomposition

Your current platform won't show you how your portfolio ranks against real investors in the same risk category. Create your secure Pure Benchmarks account and see exactly where you stand.

See Your Free Benchmark Report

Frequently asked questions

Is there a way to track my portfolio decisions and know which ones were the best?

Yes, two ways. Keep a decision journal recording the thesis, horizon, disconfirming evidence and the alternative you passed on, then grade each closed position against it. Or reconstruct the counterfactual from your transaction history and measure every decision against what the portfolio would be worth had you changed nothing. The journal is better on reasoning; the counterfactual is better on evidence, works retroactively, and is what Pure Benchmarks, our own product, automates from connected holdings.

What should I record for each investment decision?

The purchase details, the thesis, the expected outcome, the time horizon, what would make the thesis wrong, and what you would have bought instead. On exit, record the return, the result relative to an appropriate benchmark, and a separate judgment on decision quality. The last field is the one that matters, because it forces you to state whether the decision worked rather than whether the price rose.

What is the difference between tracking what I own and evaluating my decisions?

Tracking what you own reports the current value and return of positions. Evaluating decisions asks whether the act of buying, selling or holding added anything over the alternative you had at that moment. A portfolio can be up while every decision inside it destroyed value relative to leaving it alone, and a tracker will never show you that.

Can Sharesight or Koyfin tell me which decisions were best?

Partly. Sharesight reports money-weighted returns and contribution analysis, so it shows which holdings drove performance and correctly accounts for the timing and size of your investments. Koyfin produces allocation and selection attribution against a model portfolio. Neither builds the version of your portfolio in which you made no changes, so neither isolates the decision itself from the market it was made in.

How do I judge a decision that lost money?

Compare it to the alternative rather than to zero. A sale that avoided a larger drawdown was a good decision that shows as a loss on the statement, and a purchase that rose less than what you would otherwise have held was a bad decision that shows as a gain. This is why the do-nothing baseline matters: it is the only comparison that makes a losing decision and a bad decision distinguishable.

Keep exploring

See how a specific firm's advisor decisions compare to other firms:

New here? Read the portfolio benchmarking FAQ or the Pure Benchmarks blog.

This page is an information baseline for comparison only. It is not investment advice and not a recommendation to buy, sell, replace, or transfer any specific asset, account, or firm. Past performance does not guarantee future results. All figures shown are illustrative.