Extending a 1930s Semi

What is specific to this house type

Extensions in general are covered elsewhere on this site. What follows is the part that is particular to the 1930s semi, because these houses were built to a small number of repeated plans and they throw up the same questions every time.

• An existing rear addition. Nearly all of them have a back projection containing the kitchen, and often an outhouse or bathroom beyond it. It is frequently half-brick, uninsulated, on shallow foundations and with a lean-to roof.
• A hipped roof, which slopes on three or four sides. That geometry makes a two-storey rear extension harder to roof than on a gable-ended house.
• An attached neighbour, so the party wall is a live issue on almost every job.
• A narrow side passage between the semi and the boundary, which constrains both access for the build and what a side extension can be.
• Early cavity walls with corrodible ties, which need checking before anything is tied into them.
• A bay window at the front and often the rear, carried on a lintel that was adequate as built.
• A generous rear garden by modern standards, which is why these houses extend well.
• Distinctive brick and detailing that a new elevation will not reproduce.

The good news is that the house type extends well. The garden depth is usually there, the plan is simple, and the result — a wide open kitchen and dining space across the back — suits how people want to live. The work is in handling the four constraints above rather than in the extension being difficult.

The existing rear addition: three options

The fourth row is the one to be wary of. Extending off the end of an existing narrow rear addition produces a long, thin, poorly lit space, and it is chosen because it looks cheaper on the drawing. Removing the addition and building the full width usually costs a little more and produces a far better house.

Matching the brick: it will not match

This is worth stating plainly because every client hopes otherwise and nobody is ever told early enough.

The brick on a 1930s house has eighty or ninety years of weathering on it. Even where the original brick type is still made — and often it is not — a new brick arrives clean, sharp-edged and a different shade. Within a few years it tones in somewhat; it never becomes indistinguishable.

So the choice is not between matching and not matching. It is between a poor attempt at matching and a deliberate design decision. The deliberate options:

1. Get as close as possible and accept the difference. A brick-matching service against a sample is worth using. Take the sample from a sheltered area rather than a weathered one, and ask for the match in daylight.
2. Use reclaimed brick of the right era, which can match well. More expensive, variable in supply, and worth sourcing before the design is finalised.
3. Change material deliberately — render, timber cladding, a dark brick, or large areas of glazing. A new element that reads as new looks considered; a near-miss brick reads as a mistake.
4. Break the junction. A recessed glazed link, a change of plane or a rainwater pipe at the junction gives the eye a line to read the change against.
5. Match the mortar, not just the brick. The mortar is a large proportion of a wall's appearance and the colour and the joint profile matter as much as the brick does. This is the detail most often got wrong.
6. Render the whole elevation so old and new read as one. A larger job and the most complete answer where matching matters.

Point five deserves emphasis. We have seen good brick matches undone by a grey modern mortar against the original's softer, sandier joint, and the wall reads as obviously new despite the brick being right.

The hipped roof problem

On a gable-ended house a two-storey rear extension is relatively straightforward: the new roof runs back off a flat gable wall.

On a hipped semi the roof slopes away on the side as well as the rear, so there is no flat plane to build back from. That has consequences.

For a single-storey extension, little changes. The new roof — flat, lean-to or with rooflights — attaches below the first floor windows and the hip is not involved.

For a two-storey extension, the hip is the problem. The options are a new hipped roof over the extension, which is more complex and loses internal headroom at the edges; a flat roof, which is simpler and must be detailed well; or converting the hip to a gable first, which gives a flat plane to build against and incidentally makes the loft far more usable.

That last option is worth taking seriously. A hip-to-gable conversion is a known quantity on this house type, it opens up the loft for a future conversion, and it simplifies the extension roof. Doing both together shares the scaffold and the roofing work.

Planning is also a factor. A two-storey rear extension on a semi is more likely to need permission than a single storey, and the effect on the attached neighbour's light and outlook is a material consideration. The hip itself is often locally characteristic, so changing it can attract comment in some streets.

The party wall, and the neighbour

On a semi this is not an edge case. Almost every extension here engages the Party Wall etc. Act in at least one respect.

The triggers: cutting into the shared wall to insert a beam or flashing, building up to or astride the boundary line, raising the party wall, and excavating within three metres of the neighbour's structure to a depth below their foundations. A rear extension on a semi commonly does two or three of those.

The process is notice, then the neighbour consents in writing or dissents. On dissent, surveyors are appointed and an award is produced setting out the work, the condition of the neighbour's property beforehand, and access rights. It costs money and takes weeks.

The practical advice is to start early and to talk to the neighbour before any notice lands. A notice arriving unannounced from a surveyor reads as a legal threat; the same information from the person next door, with the drawings, usually does not. Consent in writing avoids the surveyor route entirely and is the common outcome where the relationship is good.

A schedule of condition is worth having even where the neighbour consents. Photographs of their property before work starts settle any later question about whether a crack was there already, and it protects both sides.

Also worth considering: the neighbour may want to extend too. Where two semis extend together the shared wall is built once, the scaffold is shared, and both houses end up better. It does not often happen, and it is worth asking.

Mistakes Coventry homeowners keep making

• Extending off the end of the existing rear addition, producing a long dark corridor because it looked cheaper on the plan.
• Assuming the brick will match, then being disappointed by a wall nobody designed.
• Matching the brick and ignoring the mortar, which undoes the match.
• Tying into a cavity wall with failed ties. Check them before the new structure relies on the old one.
• Serving party wall notice late, which can delay a start by weeks.
• Forcing a two-storey extension onto a hipped roof without considering a hip-to-gable instead.
• Blocking the side passage with a side extension and losing the only route to the rear garden.
• Not planning for the loft. If a conversion is likely later, the extension and the stair position should allow for it now.

What it costs

These are the extension figures published across this site:

Removing the existing rear addition rather than building around it adds demolition and a larger footprint, and it is usually the better value of the two. Party wall surveyor fees and brick matching or reclaimed brick are separate costs worth establishing before committing.

Frequently asked questions