Investing in property refurbishments: the investor's guide

Buying a property, refurbishing it and selling it is the oldest real estate strategy there is, and also the worst explained. Online it gets told as a formula — "buy cheap, do it up, sell high" — and in practice it is an industrial project with a budget, a schedule and a supply chain, where profit doesn't appear because the market rose but because somebody delivered the works properly. This guide is written from the point of view of the investor who puts up capital but doesn't pick up a trowel: what happens inside the deal, where every euro goes, where it breaks, and what to demand before signing. With the numbers from 49 buy-improve-sell operations already settled.

In a refurbishment deal, profit is not the gap between what the flat cost and what it sold for. It's the gap between the sale price and everything it took to get there: purchase, acquisition taxes, works, professional fees, permits, months of local rates and service charges, selling costs and exit taxes. That distinction isn't accounting pedantry — it's what separates a profitable deal from one that merely looks profitable.

And there's a second distinction that matters even more: a refurbishment margin has two possible sources, and only one of them is under anyone's control.

• Market appreciation: the property is worth more at exit because prices rose during the works. It's a gift when it happens, but nobody controls it, and an investment thesis that leans on it is a bet, not a project.
• Value created on the asset: the property is worth more because it is better. The layout has changed and gained a bedroom; services at the end of their life have been replaced; a habitability certificate has been resolved; the energy rating has moved up, which already feeds into the sale price. That step up in category can be budgeted, and can be audited before you commit.

A well-structured deal stands entirely on the second source and treats the first as a bonus. When the estimated return on a fact sheet only works if you assume prices rise X% over the coming months, the analysis has already told you everything you need to know.

This is the part almost nobody shows you. The works themselves — the only thing people associate with the word "refurbishment" — are rarely the biggest line: in a typical deal they weigh considerably less than the purchase, and often less than taxes, fees and selling costs combined. The ranges below are indicative, expressed against the total cost of a residential buy-refurbish-sell deal, and vary widely by market, building age and scope.

Three practical conclusions. First: the margin is won at purchase, because the dominant line is the only one that can no longer be touched afterwards. Second: acquisition and exit costs together take something like 10–17% of total cost and are fixed and certain, while appreciation is uncertain — which is why a deal needs a wide spread to be viable at all. Third: one line, holding costs, grows on its own with the calendar. Which brings us to the heart of this article.

Invernova publishes its full record of closed operations: 49 settled deals and 0 failed, €13.2M of capital managed, €10.7M returned to investors and €3.03M of profit generated, at an average achieved return of 19.6%. These are finished results, not projections. And the most useful thing in that record isn't the average — it's what happens when you cross return with term.

Read that table top to bottom and something uncomfortable shows up: raw return and annual return do not move together. The best headline figure in the table — 52.2% — took three years, and per year it trails an 18.4% closed in five months. A 38.3% that looks impressive on paper is, per year, one of the worst on the list. That unremarkable-looking 18.4% is the third-best deal per unit of time. The "annual equivalent" in the last column is simply the return divided by the months and multiplied by twelve, without compounding; it's an approximation, but it makes the point.

The aggregate numbers say the same thing. Across the 49 terms published as of October 2026, the average is around 15.1 months; strip out the four deals that ran to three years or more and it drops to roughly 13. That gap of two months in the average is produced by 4 deals out of 49: around 8% of cases run very long indeed, and that 8% is precisely the risk an investor takes on when reading "estimated term: 12 to 24 months".

After dozens of deals, the failures aren't varied: the same four repeat, and all four can be spotted before you commit if you know what to ask.

1. Buying too expensively

The most common failure and the most irreversible. Buying 5% above the real comparable — closed transactions in the same neighbourhood for equivalent product, not the portal asking price — wipes out a huge share of the margin on a twelve-month deal, and no amount of refurbishment recovers it. When there's competition for the asset and pressure to close, this is the mistake that gets made.

2. A budget with no contingency

In stock built before 1980, surprises turn up with statistical regularity: structure, services at their limit, damp, courtyards, protected features. A works budget without an explicit contingency line — 10% to 15% is prudent in older buildings — isn't a tight budget, it's an incomplete one. And the follow-up question matters just as much: if it overruns, who absorbs it, the developer or the investor?

3. Permitting

This is where the months disappear. An intervention resolved with a prior notification is nothing like one that needs a technical project and a major works licence, and in cities with heavy administrative load the time between applying and being able to start is a variable in its own right. A schedule that doesn't separate permitting months from construction months is incomplete, and it's the number one reason twelve months become eighteen.

4. The exit

A refurbishment only becomes money when somebody buys. And the buyer doesn't pay what the works cost: they pay what the local comparable says refurbished product is worth. Hence two classic errors — over-specifying a property above the price ceiling of its own street, and underestimating selling time on illiquid product, such as a house with a plot versus a two-bedroom flat in a consolidated area. We cover this with area-by-area data in the guide to real estate investment in Barcelona.

Doing it yourself requires 100% of the capital, a genuine ability to run a construction project and several months of your time. If what you want is exposure to the model without becoming a developer, there are three routes with very different profiles:

The fundamental difference between the last two columns is what you get if the deal goes better than expected. With debt you collect the agreed interest and that's that; with co-investment you share in the actual result. In exchange, with debt your figure doesn't depend on the final margin being delivered. Neither is better in the abstract: it depends on whether you prefer certainty or participation. We compare formats and market players in platforms for investing in property, and the general approach in investing in property without buying a flat.

Invernova's model is the third one: buy, improve and sell, specific identified deals of 12 to 24 months, with the company putting its own capital into each and the LTV published on the fact sheet. When the property is bought at a discount at auction the structure is the same, and we cover it in investing in property auctions without bidding. The full process is on how it works.

Six questions. If a fact sheet doesn't let you answer them, the problem isn't the deal — it's the information.

1. 01What's the purchase price against the real closed comparable? Not against the portal asking price. This is the number that decides the margin.
2. 02Is the works budget fixed, and does it carry a contingency line? And if it overruns, who absorbs it?
3. 03Does the schedule separate permitting from construction? A single headline term with no breakdown hides the real calendar risk.
4. 04What's the LTV? That is, what percentage of the property's value the invested capital represents. 55% means the asset would have to fall by nearly half before the cushion disappears; 80% leaves very little room for error.
5. 05Who else is putting money in, and how much? Having the party delivering the works with their own capital inside aligns incentives better than any contractual clause.
6. 06When and how do you get paid? In buy-improve-sell there are no monthly rents: capital and profit come back when the asset is sold. If you need that money sooner, this isn't your investment.

And a word on sizing: if what you're deciding is how much of your portfolio should go into deals like these, the concrete numbers are in where to invest €50,000 and in the guide to investing in property online.

• Execution risk. The dominant risk in this model. Cost overruns, trades that fail to deliver, hidden defects in older buildings: all of it comes out of the profit before it comes out of anyone else's pocket.
• Term risk. Documented above with specific names and months: in the record, roughly one deal in ten has run to three years or more. An estimated term is an estimate, not a commitment.
• Exit market risk. The sale price is set by the comparable at the moment of selling, not the one at the moment of buying.
• Regulatory and tax risk. Permits, energy-efficiency rules and tax rates change, and they change by region. Always verify the rules in force with the Agencia Tributaria and your regional tax authority.
• Illiquidity. Capital is locked up until the property is sold. There is no secondary market and no early redemption, and the term can run long.
• Estimated returns. The historical results cited are real and public; the return of each open deal is an estimate as of the date of its card.