The fund factsheet was four pages long. Top page had the manager’s commentary, second page had the asset allocation, third page had the country breakdown, fourth page had the holdings table — twenty-five lines, each row showing position name, ISIN, sector, weight, and one-year return.
I needed page four in a spreadsheet so I could line it up against three other funds and see where I had overlap. The plan was: copy the table, paste into Excel, sort by ISIN to spot duplicates, decide whether to rebalance.
I selected the table in the PDF, hit Ctrl+C, pasted into Excel. The 25 rows landed as a single column of mashed-together text — ISINs running into sector names, weights stuck to the percentages from the next column, the one-year return numbers in the wrong cells entirely. I gave up after the third row and started retyping.
Forty minutes later, I had the data clean. I sorted by ISIN and found that three of my four funds held the same five mega-cap stocks at non-trivial weight. The whole point of holding multiple funds had been diversification. I was paying four management fees for what was effectively two funds with extra steps.
The retyping took longer than the analysis. And the analysis was the only part that actually changed my portfolio.
This is, as far as I can tell, what every retail investor who tries to do their own due diligence runs into. The good information lives inside PDFs. Getting it out of the PDFs eats up the time you’d otherwise spend thinking about what the information means.
Why investment PDFs are particularly bad
Fund houses, brokers, and listed companies all publish PDFs for the same reason — it’s the format that survives intact across email, web, and print. A PDF of a quarterly report looks the same on a phone, on a laptop, and stapled together on a desk.
What they don’t optimise for is downstream use. The fund factsheet is designed to be read. It is not designed to be processed. The fact that the holdings table is a table — with rows and columns the human eye treats as data — is presentational. Underneath, the PDF stores those rows and columns as text fragments floating at coordinates. There is no spreadsheet hidden inside the PDF. You can’t extract it cleanly without a tool that reconstructs the table from the floating fragments.
This applies to fund factsheets, quarterly reports, broker statements, annual reports, pension valuations, and ISA / SIPP statements. The tables you actually care about as an investor — the ones with numbers in them — are precisely the ones PDF copy-paste mangles worst.
Why retyping is the wrong answer (even though everyone does it)
I’m being honest about the forty minutes because I think most retail investors lose much more time to retyping than they realise. Anyone who’s tracked a portfolio in Excel for more than two years has retyped at least one annual fund report by hand.
The problem is that retyping introduces errors that are almost impossible to spot after the fact. A 4.7% weight retyped as 4.1% is a wrong number that looks plausible. You won’t catch it on a glance. You won’t catch it on the sort. You’ll catch it the next quarter when the same fund mysteriously appears to have rebalanced when it hasn’t.
The compounding problem is that you’ll trust the spreadsheet you’ve spent two evenings building, because you’ve spent two evenings on it. The effort tricks you into thinking the numbers are reliable.
The cleanest fix is not to retype. Get the table out of the PDF directly, then spend your effort on the analysis, which is the part that actually changes decisions.
What the better tools actually do
Tools that extract data from PDF to Excel read the PDF’s internal text-fragment layout, cluster the fragments by row and column based on their coordinates, and output a proper spreadsheet — one cell per data point, rows aligned, columns named.
For a fund factsheet holdings table this means uploading the PDF, getting an XLSX back with twenty-five rows and five columns mapping to the original table. The extraction isn’t always 100% perfect on weird layouts — merged header cells sometimes get split, percent signs sometimes land in their own column — but the cleanup is two minutes of light editing instead of forty minutes of typing.
I now run every fund factsheet, every quarterly report, and every broker statement through extraction before I even read them properly. It’s become a reflex. The PDF arrives, I extract, I look at the spreadsheet. The PDF is for reference if I need to re-check something against the original.
The specific tables I extract and what I do with them
Holdings tables from fund factsheets. I keep a rolling spreadsheet of every fund I own with a tab per fund showing top 20 holdings each quarter. When I add a new fund, the first check is overlap with what I already hold. The diversification claim of “I have four equity funds” collapses if all four funds hold the same fifteen mega-caps.
Financial summaries from annual reports. Five-year revenue, EBIT, free cash flow, net debt. The PDF has these as a table on a single page; extracted to Excel, I can graph them and spot the company that grew revenue 10% a year by adding debt.
Broker dividend statements. For tax year reporting. UK brokers issue these as PDFs. Extracting them to Excel lets me sum dividends received across multiple accounts and reconcile against my own tracker.
Pension valuations. Annual statements from workplace pensions. Extracted, I can build a real allocation across all my retirement accounts rather than mentally tracking three separate platforms.
The mistakes the tool doesn’t fix
Extraction is not a guarantee of clean data. It just gets you to the data faster. The investor mistakes that lose real money happen after the data is in Excel: ignoring footnotes, misaligning fiscal calendars, trusting top 10 holdings snapshots without context, and forgetting currency conversion when comparing US-listed and UK-listed positions.
None of these are problems extraction can solve. They’re investor literacy problems. The tool removes the friction. Your brain still has to do the work.
If you’re advising others on this stuff
If you’ve gotten good enough at portfolio analysis that friends start asking, there’s a moment where the casual help becomes a side business. A handful of people I know moved from helping family with their ISA to taking small fees for portfolio reviews — usually flat-fee one-off work, sometimes regulated, sometimes not, depending on jurisdiction.
If you go that direction, you’ll need a way for clients to find you. The standard route is a small website plus a few reputable listings — and the directories that work for independent financial advisers aren’t the same ones that work for, say, plumbers. Industry-specific directories matter more than general business listings, but a profile on a generalist site like Sooper Articles business section is a cheap way to list your advisory business and pick up the long-tail search traffic from people typing “where to find a fee-only adviser.” Combined with a Google Business profile and one good professional headshot, that’s most of the first-year client acquisition done.
The shortest version of this
If you’re a retail investor who actually reads the documents, you’re going to spend hours each year extracting tables from PDFs. Don’t retype. Run the file through an extraction tool, fix the two or three cells that came out wrong, spend the rest of the time on the analysis. The analysis is the only part that matters. Everything before it is logistics.
